243 40 8MB
English Pages XV, 194 str.: ilustr.; 25 cm [206] Year 2016;2017
Springer Series in Materials Science 249
Ji-Guang Zhang Wu Xu Wesley A. Henderson
Lithium Metal Anodes and Rechargeable Lithium Metal Batteries
Springer Series in Materials Science Volume 249
Series editors Robert Hull, Charlottesville, USA Chennupati Jagadish, Canberra, Australia Yoshiyuki Kawazoe, Sendai, Japan Richard M. Osgood, New York, USA Jürgen Parisi, Oldenburg, Germany Tae-Yeon Seong, Seoul, Korea Shin-ichi Uchida, Tokyo, Japan Zhiming M. Wang, Chengdu, China
The Springer Series in Materials Science covers the complete spectrum of materials physics, including fundamental principles, physical properties, materials theory and design. Recognizing the increasing importance of materials science in future device technologies, the book titles in this series reflect the state-of-the-art in understanding and controlling the structure and properties of all important classes of materials.
More information about this series at http://www.springer.com/series/856
Ji-Guang Zhang Wu Xu Wesley A. Henderson •
Lithium Metal Anodes and Rechargeable Lithium Metal Batteries
123
Ji-Guang Zhang Energy and Environment Directorate Pacific Northwest National Laboratory Richland, WA USA
Wesley A. Henderson U.S. Army Research Office (ARO) Research Triangle Park, NC USA
Wu Xu Energy and Environment Directorate Pacific Northwest National Laboratory Richland, WA USA
ISSN 0933-033X ISSN 2196-2812 (electronic) Springer Series in Materials Science ISBN 978-3-319-44053-8 ISBN 978-3-319-44054-5 (eBook) DOI 10.1007/978-3-319-44054-5 Library of Congress Control Number: 2016947771 © Springer International Publishing Switzerland 2017 The views and opinions expressed are those of the authors and do not necessarily reflect those of any agency of the U.S. Government. This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG Switzerland
Preface
This book provides an overview of the extensive research spanning more than four decades in the understanding and utilization of lithium (Li) metal anodes for rechargeable Li-metal batteries with a particular emphasis on the barriers, possible solutions, and potential applications in this important field. It may be served as a basic reference for readers interested in contributing to the further advancement of Li anodes and rechargeable Li-metal batteries. We wish to express our sincere appreciation to all of the collaborators who have participated in our Li metal battery research. This work was supported by the Joint Center for Energy Storage Research (JCESR), an Energy Innovation Hub funded by the U.S. Department of Energy (DOE), Office of Science, Basic Energy Sciences and by the Advanced Batteries Materials Research (BMR) Program funded by the Assistant Secretary for Energy Efficiency and Renewable Energy (EERE), Office of Vehicle Technology of the DOE. Richland, USA Richland, USA Research Triangle Park, USA
Ji-Guang Zhang Wu Xu Wesley A. Henderson
vii
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1 3
2 Characterization and Modeling of Lithium Dendrite Growth . . . . . . 2.1 Characterization of Lithium Dendrite Growth . . . . . . . . . . . . . . . . 2.1.1 Characterization of Surface Morphologies . . . . . . . . . . . . 2.1.2 Characterization Methods for Surface Chemistry . . . . . . . 2.1.3 Other Characterization Techniques . . . . . . . . . . . . . . . . . 2.2 Effect of SEI Layer on Lithium Dendrite Growth . . . . . . . . . . . . . 2.2.1 “Dead” Lithium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Interphasial Layer and Formation of Mossy Lithium . . . . 2.3 Modeling of Lithium Dendrite Growth . . . . . . . . . . . . . . . . . . . . . 2.3.1 General Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2 Effect of Current Density. . . . . . . . . . . . . . . . . . . . . . . . . 2.3.3 Importance of Interfacial Elastic Strength . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5 5 6 16 18 21 25 27 29 31 34 36 36
3 High Coulombic Efficiency of Lithium Plating/Stripping and Lithium Dendrite Prevention . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Coulombic Efficiency of Lithium Plating/Stripping . . . . . . . . 3.2 Electrolyte and In Situ Formed Solid Electrolyte Interphase . 3.2.1 Influence of Solvents . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Influence of Lithium Salts . . . . . . . . . . . . . . . . . . . . 3.2.3 Influence of Additives . . . . . . . . . . . . . . . . . . . . . . . 3.2.4 Influence of Ionic Liquids . . . . . . . . . . . . . . . . . . . . 3.2.5 Importance of Electrolyte Concentration . . . . . . . . . 3.2.6 Self-healing Electrostatic Shield Mechanism . . . . . . 3.3 Ex Situ Formed Surface Coating . . . . . . . . . . . . . . . . . . . . . 3.4 Mechanical Blocking and Solid Electrolytes . . . . . . . . . . . . . 3.4.1 Solid Polymer Electrolytes . . . . . . . . . . . . . . . . . . . 3.4.2 Solid Inorganic Electrolytes . . . . . . . . . . . . . . . . . . .
45 45 47 49 61 67 72 75 79 83 86 86 91
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
ix
x
Contents
3.5
Effect of Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.1 Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.2 Surface Layers and Underpotential Deposition/ Stripping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5.3 Surface Roughness . . . . . . . . . . . . . . . . . . . . . . . . . 3.6 Influence of Charge/Discharge Profiles . . . . . . . . . . . . . . . . . 3.6.1 Influence of Pulsed Plating . . . . . . . . . . . . . . . . . . . 3.6.2 Influence of Plated Charge . . . . . . . . . . . . . . . . . . . 3.6.3 Influence of Plating (Charge) Current Density . . . . . 3.6.4 Influence of Stripping (Discharge) Current Density . 3.7 Effect of Rest/Storage Time . . . . . . . . . . . . . . . . . . . . . . . . . 3.8 Effect of Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.9 Effect of Stack Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.... ....
96 97
. . . . . . . . . . . .
100 103 105 105 106 107 114 116 118 123 126 127
4 Application of Lithium Metal Anodes . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Lithium Metal Batteries with Lithium Intercalation Cathodes . . . . 4.2 Lithium Metal Anodes in Lithium–Sulfur Batteries . . . . . . . . . . . 4.2.1 Performance and Characteristics of Lithium–Sulfur Batteries . . . . . . . . . . . . . . . . . . . . . . . 4.2.2 High Coulombic Efficiency and Dendrite Prevention. . . . 4.3 Lithium Metal Anodes in Lithium-Air Batteries . . . . . . . . . . . . . . 4.3.1 Li-Air Batteries Using Protected Lithium Electrodes . . . . 4.3.2 Lithium-Air Batteries Using Solid Electrolytes . . . . . . . . 4.4 Anode-Free Lithium Batteries . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
153 153 156
. . . . . . . . . . . .
. . . . . . . . . . . .
. . . . . . . . . . . .
157 163 172 175 175 177 179
5 Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
About the Authors
Dr. Ji-Guang (Jason) Zhang is a Laboratory Fellow of the Pacific Northwest National Laboratory (PNNL). He is the group leader for PNNL’s efforts in energy storage for transportation applications and has 25 years of experience in the development of energy storage devices, including Li-ion batteries, Li-air batteries, Li-metal batteries, Li-S batteries, and thin-film solid-state batteries. He was the co-recipient of two R&D 100 awards, holds 18 patents (with another 17 patents pending), and has published more than 200 papers in refereed journals. He was named Thomson Reuters’ Highly Cited Researchers-2015 in the Engineering category. Dr. Wu Xu is a Senior Research Scientist in the Energy and Environment Directorate of Pacific Northwest National Laboratory. His main research interests include the development of electrolytes and electrode materials for various energy storage systems (such as lithium batteries, organic redox flow batteries and supercapacitors), the protection of lithium metal anode, and the investigation of electrode/electrolyte interfaces. He obtained his doctoral degree from the National University of Singapore in early 2000. He has had more than 120 papers published in peer-reviewed journals, six book chapters, and 25 U.S. patents granted with another 13 patents pending. Dr. Wesley A. Henderson received his Ph.D. in Materials Science and Engineering (University of Minnesota) in 2002. He was a researcher at Lawrence Berkeley National Laboratory (1995) and Los Alamos National Laboratory (1996–1997), as well as an NSF International Research Fellow at ENEA, Advanced Energy Technologies Division, Rome, Italy. From 2004 to 2013, he was an Assistant Research Professor (Chemistry) at the U.S. Naval Academy and an Associate Professor (Chemical and Biomolecular Engineering) at North Carolina State University. He joined Pacific Northwest National Laboratory as a Senior Research Scientist in 2014 and joined the U.S. Army Research Office in 2016. His research expertise is liquid and solid electrolytes for energy storage/conversion applications including improved electrolyte characterization tools and methods and the correlation of molecular-level interactions with electrolyte properties and device performance.
xi
Abbreviations
General Abbreviations AEI AES AFLB AFM AM-EFM BF CE CIP CV DOD EDX EFM EQCM FOM FTIR GC GC-MS HAADF HOMO IC IL LUMO MAS MO MRI MS NCA NDMSO NMR
Anion exchange ionomer Auger electron spectroscopy Anode-free Li battery Atomic force microscopy Amplitude-modulated electrostatic force microscopy Butyl formate Coulombic efficiency Contact ion pairs Cyclic voltammetry Depth of discharge Energy dispersive X-ray spectroscopy Electrostatic force microscopy Electrochemical quartz crystal microbalance Figure of merit Fourier transform infrared spectroscopy Glass-ceramic Gas chromatography–mass spectrometry High-angle annular dark field Highest occupied molecular orbital Ion chromatography Ionic liquid Lowest unoccupied molecular orbital Magic angle spinning Molecular orbital Magnetic resonance imaging Mass spectroscopy Nickel-cobalt-aluminum oxide Solvation number of DMSO molecules Nuclear magnetic resonance
xiii
xiv
OCP PDOS PLE Raman RPP SEI SEM SIMS SPE SPoM SS SSE TEM TMAFM TPD-MS UPD UPS XPS XRD
Abbreviations
Open circuit potential Projected density of states Protected Li electrode Raman spectroscopy Reverse pulse plating Solid electrolyte interphase Scanning electron microscopy Secondary ion mass spectrometry Solid polymer electrolyte Surface potential microscopy Stainless steel Solid-state electrolyte Transmission electron microscopy Tapping-mode atomic force microscopy Temperature-programmed decomposition mass spectroscopy Underpotential deposition Underpotential stripping X-ray photoelectron spectroscopy X-ray diffraction
Chemical Abbreviations 2MeTHF 2MF BETI BOB DEC DEE DFOB DFT DMC DME DMSO DOL EA EC Et2O Et-G1 FEC FSI G2 G3 G4 LATP LiBETI
2-methyltetrahydrofuran 2-methylfuran Bis(perfluoroethylsulfonyl)imide or N(SO2C2F5)2Bis(oxalato)borate or B[OC(=O)C(=O)O]2Diethyl carbonate 1,2-diethoxyethane (or ethylene glycol diethyl ether) Difluoro(oxalato)borate or BF2[OC(=O)C(=O)O]Density functional theory Dimethyl carbonate 1,2-dimethoxyethane Dimethyl sulfoxide 1,3-dioxolane Ethyl acetate Ethylene carbonate Diethyl ether Ethylene glycol diethyl ether (i.e. 1,2-diethoxyethane) Fluoroethylene carbonate Bis(fluorosulfonyl)imide or N(SO2F)2Diglyme or diethylene glycol dimethyl ether Triglyme or triethylene glycol dimethyl ether Tetraglyme or tetraethylene glycol dimethyl ether Lisicon-type Li1+xALxTi2−x(PO4)3 Lithium bis(perfluoroethylsulfonyl)imide
Abbreviations
LiBOB LiDFOB LiFSI LiTFSA LiTFSI MA MF PC PEO PP PTFE TEGDME TFSA TFSI THF THP VC VEC
xv
Lithium bis(oxalate)borate Lithium difluoro(oxalato)borate Lithium bis(fluorosulfonyl)imide Lithium bis(trifluoromethanesulfonyl)amide or LiN(SO2CF3)2 Lithium bis(trifluoromethanesulfonyl)amide or LiN(SO2CF3)2 Methyl acetate Methyl formate Propylene carbonate Poly(ethylene oxide) Polypropylene Polytetrafluoroethylene Tetra(ethylene glycol) dimethyl ether Bis(trifluoromethanesulfonyl)amide or N(SO2CF3)2Bis(trifluoromethanesulfonyl)imide or N(SO2CF3)2Tetrahydrofuran Tetrahydropyran Vinylene carbonate Vinylethylene carbonate
Chapter 1
Introduction
Lithium (Li) metal is an ideal anode material for rechargeable batteries due to its extremely high theoretical specific capacity (3860 mAh g−1), the lowest negative electrochemical potential (−3.040 V versus standard hydrogen electrode), and low density (0.534 g cm−3); thus rechargeable Li metal batteries have been investigated extensively during the last 40 years (Whittingham 2012; Aurbach and Cohen 1996; Xu et al. 2014). Unfortunately, rechargeable batteries based on Li metal anode have not yet been commercialized. There are two main barriers to the development of Li metal batteries. One is the growth of Li dendrites during repeated charge/discharge processes, and the other is the low Coulombic efficiency (CE) of these processes. These two barriers lead to two critical problems for Li metal anode: internal short circuits caused by dendrites—a safety hazard—and short cycle life of the battery due to reactions between Li metal and electrolyte, consumption of electrolyte, formation of inactive Li, and continuous increase of cell impedance. Although low CE can be partially compensated by an excess amount of Li—for example, an excess amount of 300 % of Li was a common solution in the early development of Li metal batteries—the dendrite growth-related battery failure, sometimes dramatic failure that led to fire and other hazards, and the emergence of Li-ion batteries have largely diminished industry’s efforts on the development of rechargeable Li metal batteries since the early 1990s. Figure 1.1a, b show the schematic diagram of a typical Li-ion battery and a Li metal battery, respectively. In Li-ion batteries, graphite has been widely used as the anode material because Li ions can be intercalated into its layered structure so dendrite growth can be largely prevented. In Li metal batteries, the cathode shown in Fig. 1.1b can be replaced by sulfur (for a Li–S battery) or air electrode (for a Li-air battery). Figure 1.1c shows the typical morphology of a Li dendrite and the major problems associated with dendrites and low CE of the Li deposition/stripping processes. With the urgent need for the “next generation” rechargeable batteries, such as Li–S batteries (Bruce et al. 2012; Ji et al. 2009), Li-air batteries (Girishkumar et al. 2010; Lee et al. 2011), as well as rechargeable Li metal batteries that use other Li © Springer International Publishing Switzerland 2017 J.-G. Zhang et al., Lithium Metal Anodes and Rechargeable Lithium Metal Batteries, Springer Series in Materials Science 249, DOI 10.1007/978-3-319-44054-5_1
1
2
1
Introduction
Fig. 1.1 Schematic diagram of a Li-ion batteries; b Li metal batteries; c The typical morphology of Li dendrite and main problems related to dendrites and low CE. Morphology picture in c is reproduced with permission, Copyright 1976, J. Crystal Growth (Chianelli 1976)
intercalation or Li conversion compounds as the cathode, the use of Li metal anodes has attracted significant interests in recent years. Over the last 40 years, Li dendrite formation has been widely analyzed (Aurbach and Cohen 1996; Chianelli 1976; Aurbach et al. 2002; Gireaud et al. 2006; Chandrashekar et al. 2012) and simulated (Monroe and Newman 2005; Tang et al. 2009; Yamaki 1998). Most approaches to dendrite prevention focus on improving the stability and uniformity of the solid electrolyte interphase (SEI) layer on Li surface by adjusting electrolyte components and optimizing SEI formation additives (Aurbach et al. 2002; Gireaud et al. 2006; Ota et al. 2004; Shiraishi et al. 1999; Lee et al. 2007). However, because metallic Li is thermodynamically unstable with organic solvents [as indicated by Aurbach et al. (2002)], it is very difficult to achieve sufficient passivation with a Li electrode in liquid solutions. As an alternative, various mechanical barriers, either ex situ coated polymer layers or inorganic solid-state blocking layers with high shear modulus, have been proposed to block dendrite penetration (Monroe and Newman 2005; Armand et al. 1981; Li et al. 2002; Balsara et al. 2008, 2009; Visco et al. 2004). These approaches rely on a strong mechanical barrier to prevent Li dendrite penetration through separator, but do not change the fundamental, self-amplifying behavior of the dendrite growth. In other words, these methods do not prevent Li dendrites from growing during long-term cycling, or barely improve the CE of Li deposition/stripping, thus they are not suitable for practical rechargeable Li batteries. Although some factors that can suppress Li dendrite growth may also lead to high CE of Li cycling, in many experimental studies, they are not always directly correlated with each other. Overall, high CE is a more fundamental criterion required for stable cycling of a Li metal anode and the related Li metal batteries. To have a high CE, side reactions between native or freshly deposited Li and electrolyte have to be minimized. These side reactions are proportional to the chemical and electrochemical activity of native Li when it is in direct contact with the surrounding electrolyte. They are also proportional to the surface area of deposited Li. Therefore, a high CE of Li cycling is usually a direct result of low reactivity between freshly deposited Li and electrolyte as well as a low surface area of the deposited Li. On the one hand, a dendritic Li deposition always has a high surface
1 Introduction
3
area. This means that the high CE of Li deposition/stripping is always related to a low surface area Li deposition and a suppressed Li dendrite growth. A stable CE value during long-term cycling also means that an SEI layer formed during Li deposition is relatively stable and very minimal formation of new SEI layers occurs during each cycle. On the other hand, some electrolytes can lead to dendrite-free Li deposition, but exhibit a CE of only less than 80 % (Qian et al. 2015; Ding et al. 2013). This phenomenon often is related to a highly aligned nanoarray of Li structure that has no Li extrusion but still exhibits a high surface area, and freshly deposited Li is highly reactive with the surrounding electrolyte during the cycling process. Therefore, the enhancement of CE is a more fundamental factor controlling long-term, stable cycling of a Li metal anode. In Chap. 2 of this book, we will first review various instruments/tools that are critical for the characterization of Li dendrite growth/stripping processes and analysis on the composition of the surface films formed during Li deposition, then we will review the general models of the dendrite growth processes. The effect of the SEI layer on the modeling of Li dendrite growth will also be discussed, which has often been neglected in the literature. In Chap. 3, various factors that affect CE of Li cycling and dendrite growth will be discussed, together with an emphasis on the enhancement of CE. This is partially due to the fact that almost all literature articles report the CE and Li deposition morphology together and a separate description will lead to significant duplications. Chapter 4 of the book will discuss the specific application of Li metal anodes in several key rechargeable Li metal batteries, including Li-air batteries, Li–S batteries, and Li metal batteries using other Li intercalation and Li conversion compounds as cathodes. Finally, a perspective on the future development and application of Li metal batteries will be discussed in Chap. 5.
References Armand MB, Duclot MJ, Rigaud P (1981) Polymer solid electrolytes: stability domain. Solid State Ionics 3–4:429–430. doi:http://dx.doi.org/10.1016/0167-2738(81)90126-0 Aurbach D, Cohen Y (1996) The application of atomic force microscopy for the study of Li deposition processes. J Electrochem Soc 143(11):3525–3532 Aurbach D, Zinigrad E, Cohen Y, Teller H (2002) A short review of failure mechanisms of lithium metal and lithiated graphite anodes in liquid electrolyte solutions. Solid State Ionics 148:405– 416 Balsara N, Panday A, Mullin SA (2008) Polymer electrolytes for high energy density lithium batteries. http://www1.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2008/exploratory_ battery/merit08_balsara.pdf. Accessed 12 Jan 2016 Balsara NP, Singh M, Eitouni HB, Gomez ED (2009) High elastic modulus polymer electrolytes. US Patent Appl No 0263725 A1 Bruce PG, Freunberger SA, Hardwick LJ, Tarascon J-M (2012) Li-O2 and Li-S batteries with high energy storage. Nat Mater 11(1):19–29
4
1
Introduction
Chandrashekar S, Trease NM, Chang HJ, Du L-S, Grey CP, Jerschow A (2012) 7Li MRI of Li batteries reveals location of microstructural lithium. Nat Mater 11(4):311–315. doi:http://www. nature.com/nmat/journal/v11/n4/abs/nmat3246.html#supplementary-information Chianelli RR (1976) Microscopic studies of transition metal chalcogenides. J Cryst Growth 34 (2):239–244 Ding F, Xu W, Graff GL, Zhang J, Sushko ML, Chen X, Shao Y, Engelhard MH, Nie Z, Xiao J, Liu X, Sushko PV, Liu J, Zhang J-G (2013) Dendrite-free lithium deposition via self-healing electrostatic shield mechanism. J Am Chem Soc 135(11):4450–4456. doi:10.1021/ja312241y Gireaud L, Grugeon S, Laruelle S, Yrieix B, Tarascon JM (2006) Lithium metal stripping/plating mechanisms studies: a metallurgical approach. Electrochem Commun 8(10):1639–1649. doi:10.1016/j.elecom.2006.07.037 Girishkumar G, McCloskey B, Luntz AC, Swanson S, Wilcke W (2010) Lithium—air battery: promise and challenges. J Phys Chem Lett 1(14):2193–2203 Ji XL, Lee KT, Nazar LF (2009) A highly ordered nanostructured carbon-sulphur cathode for lithium-sulphur batteries. Nat Mater 8:500–506 Lee YM, Seo JE, Lee Y-G, Lee SH, Cho KY, Park J-K (2007) Effects of triacetoxyvinylsilane as SEI layer additive on electrochemical performance of lithium metal secondary battery. Electrochem Solid-State Lett 10(9):A216–A219 Lee J-S, Kim ST, Cao R, Choi N-S, Liu M, Lee KT, Cho J (2011) Metal-air batteries with high energy density: Li–air versus Zn–air. Adv Energy Mater 1(1):34–50 Li Y, Fedkiw PS, Khan SA (2002) Lithium/V6O13 cells using silica nanoparticle-based composite electrolyte. Electrochimica Acta 47(24):3853–3861. doi:http://dx.doi.org/10.1016/S0013-4686 (02)00326-2 Monroe C, Newman J (2005) The impact of elastic deformation on deposition kinetics at lithium/polymer interfaces. J Electrochem Soc 152(2):A396–A404. doi:10.1149/1.1850854 Ota H, Shima K, Ue M, Yamaki J-I (2004) Effect of vinylene carbonate as additive to electrolyte for lithium metal anode. Electrochim Acta 49(4):565–572. doi:10.1016/j.electacta.2003.09.010 Qian J, Xu W, Bhattacharya P, Engelhard M, Henderson WA, Zhang Y, Zhang J-G (2015) Dendrite-free Li deposition using trace-amounts of water as an electrolyte additive. Nano Energy 15:135–144. doi:10.1016/j.nanoen.2015.04.009 Shiraishi S, Kanamura K, Takehara Z (1999) Surface condition changes in lithium metal deposited in nonaqueous electrolyte containing HF by dissolution-deposition cycles. J Electrochem Soc 146(5):1633–1639. doi:10.1149/1.1391818 Tang M, Albertus P, Newman J (2009). J Electro-chem Soc 156:A390–A399 Visco SJ, Nimon E, De Jonghe LC, Katz B, Chu MY (2004) Lithium fuel cells. In: Proceedings of the 12th international meeting on lithium batteries, Nara, Japan, June 27–July 2, 2004. The Electrochemical Society, Pennington, NJ Whittingham MS (2012) History, evolution, and future status of energy storage. Proc IEE 100 (Special Centennial Issue):1518–1534. doi:10.1109/JPROC.2012.2190170 Xu W, Wang J, Ding F, Chen X, Nasybulin E, Zhang Y, Zhang J-G (2014) Lithium metal anodes for rechargeable batteries. Energy Environ Sci 7(2):513–537. doi:10.1039/c3ee40795k Yamaki J-I (1998) J Power Sources 74:219–227
Chapter 2
Characterization and Modeling of Lithium Dendrite Growth
Dendrites are a common occurrence when electrodepositing metals. Although the term “dendrite” is prevalent throughout the scientific literature when referencing Li deposition, such structures are atypical (i.e., in general, the plated Li morphology does not consist of regular branched, tree-like structures). Instead, Li tends to plate from solution as particles/nodules or whiskers/needles/filaments, which can aggregate into more complex constructs. Due to its ubiquitous usage for Li electrochemistry, however, the term “dendrite” will be used throughout the text in this book to refer to the latter structures. It is also notable that the literature that addresses the theory of Li electrodeposition has focused largely on the numerous determinant factors that influence classical dendritic metal plating, but the evolution of the plated Li structural characteristics is in actuality dictated to a great extent by the concurrent reactions between the reactive Li and electrolyte components (i.e., SEI formation).
2.1
Characterization of Lithium Dendrite Growth
As discussed in Chap. 1, the Li dendrite growth during the Li deposition process is a critical issue for the battery safety. Extensive efforts have been made to characterize and analyze the formation and growth processes of Li dendrites in order to reveal the mechanisms of dendrite formation and growth processes and then find the approaches to suppress or prevent the dendrite formation. In the last four decades, many different characterization methods have been used to study Li electrodes and the dendrite formation, including scanning electron microscopy (SEM) (Aurbach et al. 1998; Dollé et al. 2002), optical microscopy (Howlett et al. 2003; Nishikawa et al. 2010; Arakawa et al. 1993), atomic force microscopy (AFM) (Aurbach and Cohen 1997), transmission electron microscopy (TEM) (Liu et al. 2011; Ghassemi et al. 2011), nuclear magnetic resonance (NMR) (Chang et al. 2015), Fourier transform infrared spectroscopy (FTIR) (Morigaki 2002; Aurbach et al. 1987, 1995, © Springer International Publishing Switzerland 2017 J.-G. Zhang et al., Lithium Metal Anodes and Rechargeable Lithium Metal Batteries, Springer Series in Materials Science 249, DOI 10.1007/978-3-319-44054-5_2
5
6
2 Characterization and Modeling of Lithium Dendrite Growth
1997; Kanamura et al. 1997), and x-ray photoelectron spectroscopy (XPS) (Ota et al. 2004a; Aurbach et al. 1993; Kanamura et al. 1995b). Both morphology and chemical composition of the deposited Li metal films have been extensively investigated. The main approaches used in the characterization of electrochemically deposited Li metal films are briefly introduced in this chapter.
2.1.1
Characterization of Surface Morphologies
2.1.1.1
SEM
Among various observation methods, SEM is the most common and useful technique to examine Li electrode surface morphologies. Morphology study focuses on variation of the Li surface and formation of Li dendrites. Since the 1970s, this method has been used to study Li film growth (Dey 1976) both ex situ and in situ during dendrite growing/stripping processes (Ding et al. 2013, 2014; Zhamu et al. 2012; Ryou et al. 2012; Stark et al. 2011; Aurbach et al. 1989, 1990a, 1994b; Besenhard et al. 1987; Kanamura et al. 1995a; Yamaki et al. 1998; Shiraishi et al. 1999; Ota et al. 2004a; Gireaud et al. 2006; López et al. 2009; Nazri and Muller 1985b; Yoshimatsu et al. 1988; Choi et al. 2011; Lee et al. 2006; Yoon et al. 2008; Li et al. 2014a; Bieker et al. 2015). With its high resolution images, SEM is a powerful technique to analyze the surface change of Li metal during the deposition and stripping cycles. The effects of solvents (Aurbach et al. 1990a, b; Besenhard et al. 1987; Ota et al. 2004a; Bieker et al. 2015; Naoi et al. 1999), salts (Aurbach et al. 1994; Kanamura et al. 1995a; Yang et al. 2008), additives, (Shiraishi et al. 1999; Fujieda et al. 1994; Osaka et al. 1997b; Ota et al. 2004b; Lee et al. 2007), and other treatments (Stark et al. 2011; Thompson et al. 2011; Stark et al. 2013) have been directly discovered using SEM images. With the help of SEM observations, the correlation between the surface chemistry and morphology of Li electrodes was built (Aurbach et al. 1994), and the morphological transitions on Li metal anodes during cycling were examined (López et al. 2009). In the ex situ SEM studies, the Li samples have been transferred into the SEM instrument chamber in an inert atmosphere to avoid reaction of Li metal with ambient air as described by Kohl and coworkers (Stark et al. 2011). In the in situ SEM observation of Li dendrite formation, Orsini et al. (1998, 1999) first reported using in situ SEM to observe the cross section of plastic rechargeable Li batteries using solid polymer electrolytes. They observed the accumulation of mossy Li and growth of dendritic Li at the Li/polymer electrolyte (Fig. 2.1), which was the origin of rapid interface deterioration and capacity fading. Neudecker et al. (2000) used in situ SEM to observe variation in the behavior of the anode current collector and the overlayer during Li deposition and stripping in an in situ built solid-sate thin-film battery. The cross-sectional SEM image of the battery is shown in Fig. 2.2a. During the initial charge of the battery at 4.2 V, Li was plated between the anode current collector (i.e., Cu) and the solid-state
2.1 Characterization of Lithium Dendrite Growth
7
Fig. 2.1 a Cross section of a Li battery after one charge at 1C; b surface of the Li anode of a Li battery after one charge at 1C; c Li deposit on the Li surface after one charge at 1C. The 1C rate was 2.2 mA cm−2. Reproduced with permission—Copyright 1998, Elsevier (Orsini et al. 1998)
electrolyte (i.e., lithium phosphorous oxynitride or LiPON), and lifted the Cu/LiPON overlayer but did not significantly penetrate the Cu film (Fig. 2.2b). During discharge to 3.0 V, the plated Li was stripped from the Cu substrate but the inactive Li prevented the exposed edge of the Cu/LiPON overlayer from completely settling onto the LiPON electrolyte again (Fig. 2.2c). Recently, Sagane et al. (2013) also reported use of in situ SEM to observe Li plating and stripping reactions at the LiPON/Cu interface. They found that nucleation reactions are the rate-determining step during the Li plating process, while the Li+ cation diffusivity governs the stripping process. Nagao et al. (2013) also used in situ SEM to study the Li deposition and dissolution mechanism in a bulk-type solid-state cell with a Li2S-P2S5 solid-state electrolyte (SSE). They reported that at a deposition current density over
Fig. 2.2 In situ SEM micrographs showing cross-sectional views of a Li-free thin-film battery with an in situ plated Li anode: a 1 lm LiPON overlayer/2000 Å Cu anode current collector/1.5 lm LiPON electrolyte/3.0 lm LiCoO2 cathode on quartz glass substrate prior to the initial charge, b same cell but at the end of the initial charge to 4.2 V, and c same cell but at the end of the first discharge to 3.0 V. The markers indicate a length of 1 lm. Reproduced with permission—Copyright 2000, The electrochemical society (Neudecker et al. 2000)
8
2 Characterization and Modeling of Lithium Dendrite Growth
1 mA cm−2, the local Li deposition triggered large cracks, resulting in a decrease in the reversibility of Li deposition and dissolution. On the other hand, at a low current density of 0.01 mA cm−2, Li deposition was homogeneous thus greatly reducing the occurrence of unfavorable cracks, which enables reversible deposition and dissolution. These results suggest that homogeneous Li deposition on the surface of the SSE and suppression of the growth of Li metal along the grain boundaries inside the SSE are the keys to achieve the repetitive Li deposition and stripping without deterioration of the SSE. Clearly, SEM (either ex situ or in situ) is a very useful technique to investigate morphology variations of Li electrodes during deposition and stripping in open cells (Li et al. 2014a). However, because of the ultrahigh vacuum condition in SEM tests, most in situ SEM investigations of Li deposition/stripping processes are performed on batteries using solid polymer or inorganic SSE. It is still difficult to conduct an in situ observation of Li formation and growth in conventional liquid electrolytes, which is more relevant to most of practical applications.
2.1.1.2
Optical Microscopy
Optical microscopy is another way to observe the Li dendrites and it has been widely used as an in situ method to observe and record the Li dendrite growing/stripping processes under working conditions close to those of practical applications (Osaka et al. 1997a). Although the resolution of optical microscopy is not as high as that of SEM, it still could easily and instantaneously distinguish the surface change and dendrite formation. It is an intuitional observation on Li electrodes and very helpful to understand a continuous dendrite growing/stripping process. With digital recording devices, the dendrite formation process can be recorded as a video. Therefore, the optical microscopy technique has been widely used to analyze Li electrodes in situ. However, a special optical cell is needed for in situ optical study of Li dendritic growth. Usually, such a cell is homemade as described by Brissot et al. (1998). The optical cell could work as an airtight electrochemical cell, and the Li electrode surface could be observed by optical microscopy. In the in situ observation, optical microscopy is more often used to study the cross section of a Li electrode to observe the dendrite growth perpendicular to the surface of the Li electrode (Stark et al. 2013; Brissot et al. 1998, 1999b; Sano et al. 2011; Howlett et al. 2003; Hernandez-Maya et al. 2015). Figure 2.3 shows typical dendrites formed at different current densities during in situ optical microscopy study. At low current density (0.2 mA cm−2), needle-like and particle-like dendrites are observed, while at higher current densities ( 0.7 mA cm−2), dendrites have a tree-like or bush-like shape. Figure 2.4 shows the evolution of Li dendrites observed in the inter-electrode space while the cell is being polarized (at 0.05 mA cm−2). The needle-like dendrite grows on the negative electrode and finally contacts the positive electrode, causing short circuit of the cell.
2.1 Characterization of Lithium Dendrite Growth
9
Fig. 2.3 Typical dendrites obtained with different current densities: a J = 0.2 mA cm−2 (needle-like dendrites), b J = 0.7 mA cm−2 (tree-like dendrites), c J = 1.3 mA cm−2 (bush-like dendrites). Reproduced with permission—Copyright 1998, Elsevier (Brissot et al. 1998)
Fig. 2.4 Time variation of the dendrites observed in the inter-electrode space while polarizing the cell with J = 0.05 mA cm−2. Dendrites are seen to be needle-like. It shows that the dendrite grows on negative electrode with time and finally contacts the positive electrode, causing short circuit of the cell. Reproduced with permission —Copyright 1999, Elsevier (Brissot et al. 1999b)
2.1.1.3
AFM
AFM is another useful technique to investigate Li electrode morphology. The resolution of AFM is much better than that of optical microscopy. At the same time, AFM can give a three-dimensional (3D) morphology image that is difficult to get
10
2 Characterization and Modeling of Lithium Dendrite Growth
Fig. 2.5 AFM images (1 1 lm) of a Li electrode in a 0.5 M LiAsF6/PC solution. a an image obtained after Li deposition, 0.41 C cm−2. New Li deposits are marked by a circle. b an image of the same area after consecutive Li dissolution (0.41 C cm−2). The same area marked in a is also circled here. Reproduced with permission—Copyright 2000, American Chemical Society (Cohen et al. 2000)
from SEM or optical microscopy. In 1996, Aurbach and Cohen first used AFM to study the Li deposition processes in nonaqueous electrolyte systems (Aurbach and Cohen 1996). In that work, the basic electrochemical cell was modified to hold the highly sensitive electrodes and electrolyte solution and to isolate them from atmospheric contaminants. They found that the AFM scanning is not destructive and does not change the morphology on the surface. After that, more work (Morigaki and Ohta 1998; Aurbach et al. 1999; Cohen et al. 2000; Morigaki 2002; Mogi et al. 2002a, b, c) using in situ AFM was conducted. With the special 3D morphology of AFM images, the swelling and shrinking of Li surfaces during Li deposition and stripping processes have been discovered (Morigaki 2002). Figure 2.5 shows AFM images of a Li surface film deposited in a 0.5 M LiAsF6/ propylene carbonate (PC) electrolyte, where (a) a bump after Li deposition and (b) shrinkage after consecutive Li stripping are clearly seen (Morigaki 2002). Moreover, investigation by AFM has revealed that the structure of the Li surface consists of grain boundaries, ridge lines, and flat areas (Morigaki and Ohta 1998), which cannot be proven by other morphology test methods including SEM and optical microscopy. Based on these findings, the breakdown and reparation of the SEI films on Li electrodes during Li deposition/stripping cycles have been proposed (Fig. 2.6) and probed (Aurbach and Cohen 1996; Cohen et al. 2000). Several modified AFM methods have also been used in the characterization of Li film deposition. Shiraishi et al. reported using in situ fluid tapping-mode AFM (TMAFM) coupled with an electrochemical quartz crystal microbalance to investigate the electrochemical stripping behavior of Li metal in nonaqueous electrolytes containing a small amount of HF (Shiraishi and Kanamura 1998), and also using TMAFM with surface potential microscopy (SPoM) to study the surface condition of the electrodeposited Li on a Ni substrate (Shiraishi et al. 2001). Recently, Zhang et al. (2014) used amplitude-modulated electrostatic force microscopy (AM-EFM) (a special type of AFM) to study the surfaces of deposited Li films. The Li films
2.1 Characterization of Lithium Dendrite Growth
11
Fig. 2.6 A description of the morphology and failure mechanisms of Li electrodes during Li deposition and dissolution illustrating selected phenomena: the beginning of dendrite formation (top) and nonuniform Li dissolution accompanied by breakdown and reparation of the surface films (bottom). Reproduced with permission—Copyright 2000, American Chemical Society (Cohen et al. 2000)
were electrodeposited for 15 h on Cu foils in electrolytes of 1 M LiPF6-PC without and with 0.05 M CsPF6 additive. Li dendrites were formed in the control electrolyte (i.e., without Cs+ additive) and a smooth Li film was obtained in the Cs+ containing electrolyte. As shown in Fig. 2.7, the EFM images recorded at a probe voltage of −1 V for the Li film formed in the control electrolyte exhibit wide color variations or a strong contrast, which indicates a large fluctuation and nonuniform distribution of electric field across the detected Li surface. In comparison, the EFM image for the Li film formed in the Cs+-containing electrolyte shows narrow color variations or relatively small contrast, indicating a uniform distribution of the electric field across the Li surface and is consistent with a smooth Li film.
2.1.1.4
TEM
Due to the success of the application of optical microscopy in the observation of morphologies of deposited Li films in microscale, another electron microscopy, i.e., in situ TEM, was recently used to observe in real time the formation of Li fibers or Li dendrite growth at nanoscale (Liu et al. 2011; Ghassemi et al. 2011). Huang and coworkers first reported the direct observation of Li fiber growth on different nanowire anodes (such as silicon or tin oxide) during in situ charging of nanoscale Li-ion batteries inside a TEM (Huang et al. 2010). Li fibers of up to 35 µm long
12
2 Characterization and Modeling of Lithium Dendrite Growth Control
Cs+- containing
Topography
1.5 μm
0 μm
EFM
1.0 V
0V
Fig. 2.7 Topography and EFM images (recorded at a probe voltage of −1 V and frequency of 22 kHz) of a dried Li film after deposition from the control electrolyte without Cs+-additive (i.e., 1.0 M LiPF6/PC) and the Cs+-containing electrolyte (i.e., 1.0 M LiPF6-PC with 0.05 M CsPF6). The EFM images show the distribution of electric field across the Li surface. The wide color variations in the EFM images obtained in the control electrolyte indicate a nonuniform distribution, while the narrow color variations exhibited in the images obtained in Cs+-containing electrolyte indicate a more uniform distribution of electric field across the detected Li surface. Reproduced with permission—Copyright 2015, American Chemical Society (Zhang et al. 2014)
grew on nanowire tips along the nanowire axis in an ionic liquid (IL)-based electrolyte. After that, Yassar and coworkers also used this in situ TEM technology to study the growth of Li dendrites (Ghassemi et al. 2011). They reported clear observation of nucleation of Li+ cations at the anode/electrolyte interface and then growth of Li fibers or Li dendrites on the anode surface in a nanoscale Li-ion battery (Fig. 2.8). In situ TEM is a very promising method to observe and study Li dendrite growth in situ during the continuous charging/discharging processes of a battery, especially at the nanoscale. Although in situ TEM images have been used to reveal the formation and growth of Li dendrites during the continuous charging/discharging processes of a battery, an IL
2.1 Characterization of Lithium Dendrite Growth
13
Fig. 2.8 a Black arrows indicate an individual silicon nanorod surrounded by ionic liquid. b Arrows indicate the formation of Li islands on the nanorod. c Represents the growth of Li fibers. d The formation of kinks and growth of Li fibers are marked by black arrows. Reproduced with permission—Copyright 2011, American Institute of Physics (Ghassemi et al. 2011)
electrolyte or an SSE has to be used in most in situ TEM studies because the high vapor pressure of a practical liquid electrolyte is not compatible with the high vacuum required by a conventional TEM system. It is well known that an SEI layer formed on the surface of a Li film or dendrite is critical for Li deposition/stripping. However, the SEI formed in IL or SSE is greatly different from those formed in the conventional electrolytes used in Li metal batteries. Therefore, the interaction between Li dendrites and a practical liquid electrolyte still cannot be observed in these in situ TEM studies. Very recently, with the development of liquid cells for in situ TEM techniques, a true operando TEM investigation on Li dendrite growth has been performed in electrochemical cells with conventional liquid electrolytes for Li-ion batteries (Gu et al. 2013; Mehdi et al. 2014, 2015; Sacci et al. 2014). Sacci et al. (2014) reported the direct visualization of an initial dendritic SEI formation prior to Li deposition, and this dendritic morphology remained on the surface after Li dissolution during the in situ electrochemical TEM study, which used 1.2 M LiPF6 in ethylene carbonate (EC)/ dimethyl carbonate (DMC) (3:7 by wt) as electrolyte. Mehdi et al. (2015) used in situ electrochemical scanning TEM (STEM) to study the initial stages of SEI formation and Li dendrite evolution at the anode/electrolyte interface during Li deposition/stripping processes in 1 M LiPF6/PC electrolyte. The high-angle annular dark field (HAADF) STEM images of the anode/electrolyte interface during the first three charge-discharge cycles of this operando Li battery are shown in Fig. 2.9. The deposition and stripping of Li is clearly observed. Some electrochemically inactive or “dead” Li residues around the electrode after Li stripping for all the three cycles are present, which are no longer attached to the Pt electrode. Presently, the liquid cell TEM gives relatively lower resolution images than the open cell (i.e., vacuum conditioned) TEM does due to the limitation of cell thickness required to hold liquid electrolyte. Therefore, future improvement is required on the liquid cell fabrication (including electrodes with different alloying performance and control of the thickness) along with use of faster imaging methods. Such improvements should enable the clear observations of the initial stages of different mechanisms to be quantified on the nanometer to atomic scale. When
14
2 Characterization and Modeling of Lithium Dendrite Growth
Fig. 2.9 HAADF images of Li deposition and dissolution at the interface between the Pt working electrode and the LiPF6-PC electrolyte during the a first, b second, and c third charge/discharge cycles of the operando cell. The formation of the SEI layer (ring of contrast around the electrode), alloy formation due to Li+ ion insertion, and the presence of “dead Li” detached from the electrode can all be seen in the images at the end of the cycle, thereby demonstrating the degree of irreversibility associated with this process. Reproduced with permission—Copyright 2015, American Chemical Society (Mehdi et al. 2015)
coupled with different electrolyte compositions (i.e., solvents, salts, and additives), the improved liquid cell TEM technology may provide critical insights into the complex interfacial reactions for future Li-based and other next-generation energy storage systems.
2.1.1.5
NMR
NMR is a powerful tool for detecting chemical bonds or atomic surroundings. Recently, Bhattacharyya and Grey et al. proposed using the difference between NMR signal intensities of bulk and porous Li to identify the Li dendrite growth (Bhattacharyya et al. 2010). They have successfully used this method as an in situ tool to quantitatively observe the formation of Li dendrites in different electrolytes. Chandrashekar et al. (2012) reported using 7Li magnetic resonance imaging (MRI) technique to detect in situ the variation of Li electrode morphologies during charge and discharge processes of a symmetric Li metal cell. The 7Li NMR spectra of the Li metal resonance before (pristine) and after applying a current (charged) indicated that the area of the spectrum in the charged state was 2.3 times larger than that in the pristine state. This increase could be attributed to the formation of
2.1 Characterization of Lithium Dendrite Growth
15
Fig. 2.10 7Li 2D MRI x-y images (frequency encoding in x and phase encoding in y) in the states of pristine (a) and after charging (b), and the related SEM images of Li anode in pristine (c) and charged (d) states. Reproduced with permission—Copyright 2012, Macmillan Publishers (Chandrashekar et al. 2012)
dendritic, mossy, and other microstructural metallic Li during charging. The two-dimensional 7Li MRI images before and after the cell charging are depicted in Figs. 2.10a, b, where the cumulative signals were projected along the z direction which is perpendicular to the substrate. The MRI image of the charged battery revealed the negative electrode had a significant increase in signal of almost double, while the positive electrode showed a decrease in signal of about 23 % after charging. It indicated the location and change of microstructural Li morphology, which is consistent with findings from SEM images (Figs. 2.10c, d). Recently, Arai et al. (2015) used in situ solid-state 7Li NMR to observe Li metal deposition during overcharge in Li-ion batteries. Hu et al. also used in situ NMR and computational modeling to investigate the role of Cs+ additive (Hu et al. 2016). These works indicate that NMR not only can detect the morphology variation during the Li metal deposition process, but also can reveal the possible composition of SEI layers formed on the surfaces of Li films during the electrode plating process. By combining NMR and other characterization techniques, a more comprehensive understanding of the electrode plated Li films can be obtained.
16
2 Characterization and Modeling of Lithium Dendrite Growth
2.1.2
Characterization Methods for Surface Chemistry
2.1.2.1
FTIR
The characterization methods discussed in the previous sections mainly provide information on the morphology variations of Li depositions. Several other methods have been used to analyze the chemical compositions of the surface films formed on the surface of deposited Li. The chemical composition of the Li surface film is strongly related to the electrolyte components. In turn, the SEI film formed on the surface of a Li deposition strongly affects the Li morphology and cycling performance of a Li metal battery (Aurbach et al. 1994). In this aspect, FTIR and XPS are widely used in this field to analyze the Li surface chemistry; FTIR is more suitable for detecting the organic components, while XPS gives more information about the inorganic components. Since the 1980s, FTIR has been widely used to analyze the Li surface as a nondestructive method (Morigaki 2002; Aurbach et al. 1987, 1995, 1997; Kanamura et al. 1997). In the early years, FTIR was used as an ex situ method (Aurbach et al. 1987), but it was later developed as an in situ technique to analyze Li films during electrochemical processes (Morigaki 2002; Kanamura et al. 1997). FTIR has been used to investigate the influences of electrolyte solvents, salts, additives, and other contaminants on the Li surface. From the locations and strengths of the peaks in FTIR spectra, different chemical bonds or components on the Li electrode surface could be identified. An example is shown in Fig. 2.11 which shows FTIR spectra of Li electrodes prepared and stored for three days in EC-DMC solutions of 1 M LiAsF6, LiPF6, or LiBF4 as indicated. A spectrum of a Li electrode prepared and stored in DMC containing 0.1 M CH3OH is also presented for comparison (Aurbach et al. 1997).
2.1.2.2
XPS
It should be noted that although the FTIR technique is very useful to identify the surface components, it is limited in that it detects only the IR-active species and it cannot give the relative importance of each surface component and composition that affect the Li deposition morphology and battery performance. Therefore, other surface characterization methods and technologies besides FTIR are needed to get more detailed surface chemistry data on Li anodes. As mentioned above, XPS is another very useful tool to analyze the surface chemistry of Li electrodes; in particular, it gives more information about the elemental or inorganic components— data that FTIR cannot detect. Normally, as indicated from XPS and FTIR data by Aurbach et al. (1987), the major species in a Li surface film include Li2O, LiOH, LiF, Li2CO3, lithium alkylcarbonate (RCOOLi), and hydrocarbon. Recently,
2.1 Characterization of Lithium Dendrite Growth
17
Fig. 2.11 FTIR spectra of Li electrodes prepared and stored for three days in EC-DMC solutions of 1 M LiAsF6, LiPF6, or LiBF4 as indicated. A spectrum of a Li electrode prepared and stored in DMC containing 0.1 M CH3OH is also presented for comparison. Reproduced with permission— Copyright 1997, Elsevier (Aurbach et al. 1997)
Wenzel et al. (2015) used the in situ XPS technique to study the stability of an SSE in contact with Li metal. The key concept was to use the internal Ar ion sputter gun in a standard lab-scale photoelectron spectrometer to deposit thin metal films (e.g., Li) on the sample surface and to study the reactions between metal and SSE by photoelectron spectroscopy directly after deposition (Fig. 2.12). This approach could give information on interfacial reactions and the interfacial kinetics, especially for the interface between the alkali metal and solid electrolyte in solid-state batteries. With XPS analysis, the effects of different electrolyte solvents (Ota et al. 2004a; Aurbach et al. 1993; Kanamura et al. 1995b), lithium salts (Kanamura et al. 1995a; Fujieda et al. 1994), and additives (Shiraishi et al. 1999; Odziemkowski et al. 1992) on the Li surface chemistry have been investigated. Based on these data, especially with the vacuum etching technology of the XPS technique, not only the components but also the structural composition evolution of the SEI film can be revealed by XPS analysis (Ding et al. 2014; Kanamura et al. 1995a; Shiraishi et al. 1999; Zhang et al. 2014; Aurbach et al. 1993; Qian et al. 2015a, b).
18
2 Characterization and Modeling of Lithium Dendrite Growth
Fig. 2.12 The basic concept and setup of the in situ XPS technique. An argon ion beam is used to sputter lithium, gold or aluminum metal on the sample surface (a). In (b) the geometry is schematically shown. After deposition, the reaction products formed at the interface (d) are investigated using photoelectron spectroscopy, as shown in (c). Reproduced with permission— Copyright 2015, Elsevier (Wenzel et al. 2015)
2.1.3
Other Characterization Techniques
In addition to the methods discussed in the above sections, several other methods, including Raman spectroscopy (Raman) (Kominato et al. 1997), Auger electron spectroscopy (AES) (Ota et al. 2004a; Morigaki and Ohta 1998; Aurbach et al. 1993), and NMR (Kominato et al. 1997) have been used to analyze the surface chemistry of electroplated Li films. So far, attempts to use Raman spectroscopy to identify the surface films on the Li metal/electrolyte interphase have not been very successful. Only a few papers reported the Raman studies (Howlett et al. 2003; Rey et al. 1998a, b; Naudin et al. 2003). For example, Irish et al. used a Raman microprobe to study both in situ and ex situ the surface films formed on Li metal in contact with electrolytes of LiAsF6/tetrahydrofuran (THF) and LiAsF6/2MeTHF (Odziemkowski et al. 1992). The reaction products detected were mainly polytetrahydrofuran, some arsenolite (As2O3), and arsenious oxyfluorides F2As-O-AsF2. Raman technology might be expected to yield results similar to those of FTIR spectroscopy, but this technology is more complicated to use than FTIR (Odziemkowski et al. 1992). In addition, as indicated by Naudin et al. (2003), local heating of the samples under laser irradiation is unavoidable in Raman tests. The carbonate species on Li surface could be transformed into lithium acetylides of Li2C2 type, which gives a vibration peak of CC at about 1845 cm−1, thus giving a
2.1 Characterization of Lithium Dendrite Growth
19
faulty result to the interpretation. Therefore, the destructive effect of Raman laser beam on Li surfaces limits its use in the analysis of Li surface films. Aurbach et al. (1993) used AES to measure the Li surface after the Li was immersed in an electrolyte of 0.2 M LiAsF6/1,2-dimethoxyethane (DME) for 15 min followed by pure DME rinsing. They found that the AES spectra were similar to those seen with XPS. Carbon and oxygen were detected at the Li surface. With sputtering, the intensity of the carbon Auger peak decreased while the oxygen peak increased when compared to their initial peaks. It was suggested that the surface films of Li treated in DME consisted of two layers, the upper layer being an alkoxide film (probably LiOCH3) and the layer close to Li being a mixture of Li2O and LiOH. Kominato et al. (1997) also used AES to detect the surface compounds of Li after it was immersed in three electrolytes of EC/DMC with LiPF6, LiClO4, or LiTFSI. Except for LiF found in the Li surface film from the LiPF6-based electrolyte, all major components in the three Li surfaces were Li–O components indicating LiOH, Li2O, or other lithium oxide compounds. Morigaki and Ohta used scanning AES to analyze the Li surface in 1 M LiClO4/PC solution also used AES to detect the surface compounds of Li after it was immersed in three electrolytes of EC/DMC with LiPF6, LiClO4, or LiTFSI. Except for LiF found in the Li surface film from the LiPF6-based electrolyte, all major components in the three Li surfaces were Li–O components indicating LiOH, Li2O, or other lithium oxide compounds. Morigaki and Ohta (1998) used scanning AES to analyze the Li surface in 1 M LiClO4/PC solution. Li2CO3, Li2O, and LiOH were localized on the ridge lines and the grain boundaries of the Li surface. AES technology can provide some useful information about the Li surface components, but it is very close to that obtained from XPS analysis. In addition, the AES equipment is more difficult to access than XPS equipment, so AES analysis has been used less frequently in Li metal investigations. NMR has also been used to study the Li surface chemistry. Ota et al. (2004a, c) used NMR technology (1H, 13C and 2D spectra) to analyze the surface components of deposited Li by dissolving the surface film in anhydrous dimethyl sulfoxide (DMSO)-d6 and then recording the NMR spectra of the organic species in the DMSO-d6 solution. They found that the organic surface layer on Li metal included lithium ethoxide, lithium ethylene dicarbonate, PEO, and lithium ethylene containing an oxyethylene unit. This is an indirect method to analyze the Li surface focusing on the dissolvable organic species. The obvious limitation of this technique is the inability to analyze the insoluble inorganic compounds formed on Li surfaces. Nazri and Muller used secondary ion mass spectrometry (SIMS) to study the surface layer formed on electrochemically deposited Li on copper in a 1 M LiClO4PC electrolyte (Nazri and Muller 1985b). The obtained SIMS spectrum was complex and was difficult to interpret. Basically, the low mass range showed the fragments of PC, the salt, and water, while the high mass range indicated the presence of a polymeric material based on PC, a partially chlorinated hydrocarbon polymer, and their lithium adducts. The authors also applied the in situ x-ray diffraction (XRD) technique to the analysis of the formed Li surface films (Nazri
20
2 Characterization and Modeling of Lithium Dendrite Growth
and Muller 1985a, b, c). The presence of Li2CO3, Li2O, and polymer compounds was also detected. Temperature-programmed decomposition mass spectroscopy (TPD-MS) and gas chromatography–mass spectrometry (GC-MS) technology were also used by several groups to analyze the surface components of Li electrodes (Matsuda et al. 1995). Kominato et al. (Morigaki and Ohta 1998; Kominato et al. 1997) reported that the gases generated from Li films pretreated in EC-dimethyl carbonate (DMC) based electrolytes were mainly CH4, H2O, CO, CH3OH, CO2, and ethylene oxide. N2 was also detected if LiTFSI was used as the electrolyte salt. The GC-MS detected the same gas components. This indicated that the detected gases were generated from the organic Li compounds that were the reaction products of Li and solvents (EC and DMC) and included lithium ethylene dicarbonate and lithium methylcarbonate. Ota et al. (2004a, c) also used GC-MS to investigate the Li surface compounds generated in EC/THF electrolytes. C2H4, CO2, and C2H6 were detected and were mainly from the reductive components of EC. During TPD-MS and GC-MS measurements, the Li electrodes with the detected surface films need to be heated to give off the gases to be tested. The data from these MS measurements could provide more information on the Li surface film chemistry and support the results of other film measurements, such as FTIR and XPS. Ota et al. (2004a, c) used ion chromatography (IC) to quantitatively analyze the Li surface films. The Li films were first dissolved in high-purity water and then tested by an IC instrument. By analyzing the contents of F−, CO32−, and Li+ ions, quantitative information about the Li surface films could be obtained. They found that the Li surface film in EC-based electrolytes consisted mainly of lithium alkyl carbonate, and LiF content in the films formed in an electrolyte containing lithium imide salt was lower than those formed in the electrolytes containing LiPF6 salt. The in situ scanning vibrating electrode technique has also been used to map the surface electric field of Li electrodes (Matsuda et al. 1995; Ishikawa et al. 1997). The surface electric field on a Li electrode is based on its morphology and composition uniformity. So this technology reflects not only the surface morphology, but also the chemical composition uniformity of the Li surface. However, because the scanning step of this technology is not small enough, the definition obtained using this technology is not satisfactory. In a recent effort, Harry et al. (2014, 2015) used synchrotron hard x-ray microtomography to investigate the failure caused by dendrite growth in high-energy density, rechargeable batteries with Li metal anodes. When a symmetric Li|polymer electrolyte|Li cell was cycled at 90 °C, they found that the bulk of the dendritic structure lay within the electrode, underneath the polymer/electrode interface, during the early stage of dendrite development. Furthermore, they observed crystalline impurities, present in the uncycled Li anodes, at the base of the subsurface dendritic structures. The portion of the dendrite protruding into the electrolyte increases on cycling until it spans the electrolyte thickness, causing a short circuit. Contrary to conventional wisdom, it seems that preventing dendrite formation in polymer electrolytes depends on inhibiting the formation of subsurface structures in the Li electrode. These results demonstrate that x-ray
2.1 Characterization of Lithium Dendrite Growth
21
microtomography is another powerful tool to provide a clear failure mechanism in Li metal batteries. In summary, characterization of morphologies and surface components of electroplated Li anodes is a complicated task. SEM and FTIR/XPS are the most common methods used to investigate the surface morphology and chemistry, respectively, of the electrodeposited Li films. Many other methods discussed in this chapter also provide valuable information on Li films. However, no single technique is enough to provide comprehensive understanding of the studied Li films, especially for the surface reaction products or SEI layer formed on a Li film. Therefore, a combination of multiple characterization and analysis methods is required to have a good understanding of the properties of electrodeposited Li films.
2.2
Effect of SEI Layer on Lithium Dendrite Growth
Various models proposed in the literature have provided important guidance on the nucleation and growth of metal dendrites, especially for metals that do not react with electrolyte in a significant way. However, Li is thermodynamically unstable with any organic solvent and the two react instantaneously to form an SEI. This SEI layer will continuously break down and regrow during Li deposition/stripping processes and is critical to the real growth pattern of reactive metals such as Li. Unfortunately, most models in the literature do not consider the impact of the SEI on the Li dendrite growth mechanisms. Recently, Cheng et al. (Cheng et al. 2015) reviewed the mechanisms of SEI formation and models of SEI structure. The critical factors affecting the SEI formation, such as electrolyte component, temperature, current density, are discussed. An extensive experimental study by Steiger et al. utilizing in situ light microscopy and ex situ SEM analysis concluded, however, that growth of the whisker/needle-like structures (often termed “dendrites”) occurs as follows (Steiger et al. 2014b, 2015; Steiger 2015): • Li needle growth is generally initiated at either the substrate surface or from faceted particles of Li. • Typically, the needles grow in length, but not in breadth. • Growth often does not actually occur at the tip, but rather behind an inactive deposit located at the tip (possibly a particle of metal oxide, LiF or other SEI components) (Figs. 2.13, 2.14) (Steiger et al. 2014a). • Li+ cation transport through a thin SEI layer results in the overall needle growth which occurs—as just noted—at the needle tip (i.e., deposit-Li interface), at the base (i.e., substrate-Li interface), and at defects resulting in kinks in the needles. • Extensive shaking/twisting motions transpire during the needle growth process (Steiger et al. 2014a; Brissot et al. 1998; Nishikawa et al. 2007, 2010, 2011, 2012; Nishida et al. 2013; Yamaki et al. 1998)—a characteristic that is readily explained by the differing growth zones (kinks and interfaces) in Steiger’s model.
22
2 Characterization and Modeling of Lithium Dendrite Growth
Fig. 2.13 SEM image of Li needle deposits on W (arrows point to broadening and/or particles at tips). Reproduced with permission—Copyright 2014, Elsevier (Steiger 2015)
Fig. 2.14 Schematic describing growth of a Li needle: (green) SEI—probably mainly organic; (light blue) Li insertion areas and (red) inert tip. a Initial state before Li deposition with an inert inhomogeneity within the SEI, b after growing a straight segment by Li insertion at the substrate, c after further deposition taking place below the tip, d further deposition resulting in a kink, e additional Li inserted at the base, causing tilting motions of the whole structure and f final structure. All steps proceed by Li insertion into the growing structure. Reproduced with permission —Copyright 2014, Elsevier (Steiger et al. 2014a)
A number of these points were also previously emphasized in publications by Yamaki et al. (1998) and Nishikawa et al. (2011). Note that Steiger’s model— which explains the observed Li kinked whisker/needle-like growth patterns quite well—diverges significantly from previous models that have emphasized dendrite formation due to the depletion of Li+ cations near the electrode surface, field enhancement at the needle (“dendrite”) tips, the strong influence of concentration gradients, and stresses that induce needle motion. Figure 2.15 shows SEM images of Li deposits on a stainless steel (SS) electrode after different times during which the electrode was polarized to −150 mV (versus Li/Li+) in 1 M LiPF6-EC/DMC electrolyte and 1 M N1114TFSI-LiTFSI (IL-based) electrolytes (Stark et al. 2013).
2.2 Effect of SEI Layer on Lithium Dendrite Growth
23
Fig. 2.15 SEM images of Li deposits on a SS electrode after the indicated plating times during which the electrode was polarized to −150 mV (versus Li/Li+) in a 1 M LiPF6-EC/DMC electrolyte and b 1 M N1114TFSI-LiTFSI (IL-based) electrolytes. Reproduced with permission— Copyright 2013, The Electrochemical Society (Stark et al. 2013)
When the Li deposition rate is slow, the Li nuclei formed initially will merge together as shown in Fig. 2.15a. When the needles grow fast relative to the nucleation points, then tangled fibrous aggregates of the kinked needles tend to result (Fig. 2.15b). Very different Li deposition is sometimes observed that is referred to as nondendritic—i.e., the Li has a particulate/nodular structure, which may be fused into
24
2 Characterization and Modeling of Lithium Dendrite Growth
aggregated lumps or instead simply clustered together. This then requires a growth model that diverges from the linear needle model noted above. The Li deposition morphology is governed by Li+ cation mobility, Li0 (adatom) transport on the Li surface, and perhaps to some extent Li0 self-diffusion within the bulk of the Li. Most of the focus in many mathematical models developed is on the former (i.e., Li+ cation mobility within the electrolyte). Jäckle and Groß found that Li0 atom self-diffusion has relatively high barriers on the most energetically favorable surfaces of the Li body-centered cubic (bcc) crystal and also has a lower tendency (than Mg, for example) to adopt high-coordination configurations (Jäckle and Groß 2014). This reduces the driving force for surface reconstruction from needle-like to nodular shapes. Higher temperatures—with the corresponding increase in Li0 adatom mobility—would therefore be expected to favor more nodular Li structures (Aryanfar et al. 2015) and this will be shown below to indeed be the case. Another important consideration, however, is the Li+ cation transport rate through the SEI to the Li growth surface. As noted above, the kinked needle growth may be dictated by favorable insertion of Li0 adatoms at the tip (often below an inactive particle), the base and at defects resulting in one-dimensional growth. These may be locations where the resistance is lowest to Li+ cation transport through the SEI or alternatively locations which have the lowest interfacial energy and thus serve as a sink for the adatoms. Transport, however, will be more favorable throughout the entire SEI layer at higher temperature and/or for a more conductive/thinner SEI. Thus, temperature and electrolyte composition strongly impact the Li deposition morphology (Nishikawa et al. 2011). It will be shown below that such growth actually begins as needles which then thicken into nodules—that is, the one-dimensional, linear elongation which creates the needles transitions at some point to three-dimensional growth at the needle tips and defects or alternatively thickening of the entire needle segment(s) (Steiger et al. 2014b; Steiger 2015; Arakawa et al. 1993; Nishikawa et al. 2011, 2012). A highly aligned Li growth pattern was observed when a robust and uniform SEI layer formed on the surface of the substrate as reported recently by Qian et al. (2015a), i.e., the one-dimensional, linear elongation that creates the needles transitions at some point to three-dimensional growth. The trace amount of HF derived from the decomposition reaction of LiPF6 with H2O is electrochemically reduced during the initial Li deposition process to form a uniform and dense LiF-rich SEI layer on the surface of the substrate. This SEI layer is robust and leads to a uniform distribution of the electric field on the substrate surface. In case of low rate deposition, the merged Li particles will favor growth into linear arrays of closely packed nanorods, since neighboring nanorods constrain the formation of the kinked defects, thereby enabling uniform and dendrite-free Li deposition. The surface and cross-section images of the as-deposited, dendrite-free Li films exhibit a self-aligned and highly compact Li nanorod structure as shown in Fig. 2.16, which is consistent with a vivid blue color due to structural coloration. Similar surface morphology was also observed before by several other groups (Stark et al. 2013; Zhang et al. 2014; Qian et al. 2015a).
2.2 Effect of SEI Layer on Lithium Dendrite Growth
25
Fig. 2.16 SEM images of the morphologies of Li plated in a 1 M LiPF6-PC electrolyte with 50 ppm H2O additive: a, b surface images and c, d cross-sectional images. Reproduced with permission—Copyright 2015, Elsevier (Qian et al. 2015a)
2.2.1
“Dead” Lithium
In an early publication in 1974, the effect of aging on the cycling CE of electrodeposited Li was examined using a 1 M LiClO4-PC electrolyte (Selim and Bro 1974). It was noted that the Li was mossy in appearance and that the CE (i.e., stripping/deposition charge ratio) decreased much more rapidly with increasing age of the deposit than did the Li metal content of the deposit (as determined chemically by reacting the Li with water and monitoring the amount of evolved gas generated). For example, after 40 h the CE approached zero, while the chemical analysis indicated that 80 % of the deposited Li remained (but could not be stripped from the electrode). This electrochemically unrecoverable Li has come to be known as “dead” Li (Yoshimatsu et al. 1988; Arakawa et al. 1993; Steiger et al. 2014b). Steiger et al. proposed a mechanism for this that illustrates how the “dead” Li may remain in physical, but not electrical contact with the electrode surface (Fig. 2.17) (Steiger et al. 2014b). This explains why after stripping the electroactive Li from a
26
2 Characterization and Modeling of Lithium Dendrite Growth
Fig. 2.17 Schematic of proposed growth mechanism for mossy Li. The structure is always covered by an SEI layer: a as-deposited, b the structure of (a) after further electrodeposition. Li atoms are inserted into the metal structure. Points have been marked with black circles to illustrate that the distances between these features generally increase with Li deposition time. The large black oval shape indicates the expansion of the total structure. c The structure of (a) after a dissolution step. The tips of the structure still contain Li metal (“dead” Li), but are electrically isolated from the substrate, although still being held by the former SEI shell. d Structure of (c) after an additional electrodeposition step. The top is pushed outward by the new mossy Li growing underneath. Reproduced with permission—Copyright 2014, Elsevier (Steiger et al. 2014b)
previous electrodeposition, new needle-like Li deposits often tend to grow on the electrode surface instead of on the previous (residual) material on the electrode surface (Brissot et al. 1998). This continues until the entire surface of the electrode is covered by the needles and their SEI residue. Thus, the interphasial layer on the surface of the Li may actually grow/accumulate predominantly at the working electrode surface rather than at the interphasial layer/electrolyte interface. It has been reported from in situ visualization studies for both liquid and polymer electrolytes that dendrites are often not evident during the first few cycles at higher current densities (for which the CE is often relatively high), but then dendrites form and grow in subsequent cycles (often resulting in a continuous decline in the CE with continued cycling) (Dampier and Brummer 1977; Brissot et al. 1998, 1999b). This may be explained by the growth of short new needle-like deposits on the electrode surface during the first few cycles until the electrode is covered in its entirety by the SEI residue and “dead” Li. Then, breakaway Li dendritic structures form while Li deposits on the electrode escape in the following cycles through defects in this resistive layer, culminating in the exposure of protruding kinked needles for which Li0 adatom addition is more facile, thus resulting in rapid growth of such dendrites.
2.2 Effect of SEI Layer on Lithium Dendrite Growth
2.2.2
27
Interphasial Layer and Formation of Mossy Lithium
For Li-ion batteries in which graphite is used as the anode, the SEI is generally a thin layer of reaction products formed from the degradation of the graphite (near the electrolyte interface), Li, solvent molecules, anions and/or other electrolyte components. Comparable SEI layers may form on the Li metal surface as noted in the discussion above on the needle growth mechanism. But rather than passivating the electrode from further reactions with the electrolyte, the dead Li forms an interphasial porous degradation layer (often termed “mossy” Li) that is much more substantive than the relatively thin SEI layer formed on the graphite surface. Upon cycling, there is thus a transition from a flat, smooth morphology to a rugged structure with a surface interphasial layer that continues to grow in thickness upon cycling (Figs. 2.18, 2.19) (Naoi et al. 1996; Orsini et al. 1998, 1999; López et al. 2009, 2012; Chang et al. 2015; Lv et al. 2015). In comparison, more mossy Li is formed at low current density and more dendrite is formed at high-current density (Orsini et al. 1999). The increased resistance results in increased polarization (i.e., higher absolute voltages for plating/stripping) (Fig. 2.20) and the “dead” Li may mask portions of the electrode (i.e., reduce the active surface area available for Fig. 2.18 Sectioned Li battery after charging (0.45 mA cm−2): a first charge and b 14th charge. Reproduced with permission —Copyright 1998, Elsevier (Orsini et al. 1998)
28
2 Characterization and Modeling of Lithium Dendrite Growth
Fig. 2.19 Morphological changes during cycling between two Li electrodes in a 1 M LiPF6-PC electrolyte at a current density of 1 mA cm−2: a uncycled, b 100 cycles, c 200 cycles and d 500 cycles. Reproduced with permission—Copyright 2003, Elsevier (Howlett et al. 2003)
deposition), thus increasing the effective current density which, as will be shown below, often results in more rapid Li consumption and possibly increased dendritic Li growth. This depletion of the electrolyte and electroactive Li due to degradation
2.2 Effect of SEI Layer on Lithium Dendrite Growth
29
Fig. 2.20 Voltage profile for a Li||Li cell with a 1 M LiPF6-PC electrolyte cycled at a current density of 1 mA cm−2. Voltage spikes are due to temporary short circuits. Reproduced with permission—Copyright 2003, Elsevier (Howlett et al. 2003)
reactions, as well as the increased interfacial resistance from the porous interphasial layer, has been delineated as a principal cause of cell performance degradation and failure (Fig. 2.21) (Lv et al. 2015), although dendritic short circuiting may occur in some instances (López et al. 2009, 2012; Orsini et al. 1998, 1999; Yoshimatsu et al. 1988). Thus, the large voltage spikes in Fig. 2.20 originate from the highly resistive interphase rather than from short circuiting (which would result in a very low potential between the electrodes). Importantly, for full batteries, the loss of electrolyte (and the corresponding significant impedance increase) may ultimately be the key factor for the deterioration of the cell capacity rather than formation of the interphasial layer (Aurbach et al. 2000).
2.3
Modeling of Lithium Dendrite Growth
Significant work has been done to simulate and predict the growth pattern of Li dendrite growth in the last half-century. When an electrochemical cell containing an electrolyte with a Li salt is sufficiently polarized, Li+ cations near the negative electrode are reduced to Li metal and—depending upon the applied current density and the electrolyte’s transport properties—the Li+ cation concentration decreases resulting in anion migration toward the positive electrode until equilibrium is reached. The newly formed Li metal may deposit as a relatively compact layer or in a variety of other morphologies which are often described as dendritic. A recent review of models for dendrite initiation/propagation described three classifications (Li et al. 2014b):
30
2 Characterization and Modeling of Lithium Dendrite Growth
Fig. 2.21 Failure mechanism of the Li anode at high charge current densities: a schematic illustration of the failure mechanism and b conventional understanding of the dendrite-related failure mechanism for Li batteries. Reproduced with permission—Copyright 2014, Wiley-VCH (Lv et al. 2015)
1. Surface Tension Model—This model finds that electrodeposition is more rapid on projections rather than planar surfaces because spherical rather than linear diffusion dominates the mass transport of the active species. A larger spherical diffusion flux results in a narrowing of the growing dendrite tip; thus surface forces and mass transport dominate the kinetics governing dendrite growth (Monroe and Newman 2003, 2004, 2005; Barton and Bockris 1962; Diggle et al. 1969). 2. Brownian Statistical Simulation Model—This model focuses on the competition between ion transport and reductive deposition. When deposition probability is
2.3 Modeling of Lithium Dendrite Growth
31
low, Li+ cation transport dominates and the ions penetrate close to the substrate resulting in a dense plating morphology with a low tendency for dendrite formation. This is contrasted with the case of high-deposition probability (relative to the cation transport), which increases the rate of deposition at the tip of projecting growth structures thus increasing the tendency for dendrite formation (Mayers et al. 2012; Voss and Tomkiewicz 1985; Magan et al. 2003). 3. Chazalviel Electromigration-Limited Model—This model for dendrite growth is often cited: at a time s called the Sand time, the Li+ cation concentration drops to zero at the negative electrode with high currents, causing the potential to diverge; this results in an instability at the interface from inhomogeneities in the distribution of the surface potential, which creates a localized electric field that leads to dendrite growth. The dendrite initiation time thus corresponds to the buildup of the space charge and the dendrite propagation velocity is tied to the transport of the anions (Brissot et al. 1998; Chazalviel 1990). In addition to these, several new models have recently been proposed to describe Li plating/growth (Akolkar 2013, 2014; Chen et al. 2015; Cogswell 2015; Tang et al. 2009; Aryanfar et al. 2014). More details on a few of these important models of dendrite growth will be discussed in this section.
2.3.1
General Models
In the field of electrodeposition, metal “dendrites” are a common phenomenon. At a given electrodeposition condition, many metals, such as zinc (Zn), copper (Cu), silver (Ag), and tin (Sn), were reported to exhibit ramified morphologies (Chazalviel 1990). Fractal deposits including needle-like, snowflake-like, tree-like, bush-like, moss-like, and whisker-like structures are all referred to as dendrites in this review. Extensive work has been done on the dendrite formation mechanism during electrodeposition of Zn and Cu. Various strategies have been unitized to suppress dendrite growth in these processes (Sawada et al. 1986; Diggle et al. 1969). The reported methods to suppress Zn dendrites include special separators, alternating current or pulsed charging, and additives in the electrolytes. The last methods can be further divided into three categories: electrode structural modifiers, metallic additives, and organic additives (Lan et al. 2007). Several factors such as Zn concentration, complexing agents, anions, and additives may modify the texture and morphology of Zn electrodeposited coatings (Mendoza-Huízar et al. 2009). Recently, Aaboubi et al. (2011) investigated the effect of tartaric acid on Zn electrodeposition from a sulfate plating bath by electrochemical impedance spectroscopy (EIS), stationary polarization curves, XRD, and SEM imagery. The study shows that it is possible to obtain homogeneous, compact, and dendrite-free Zn deposits from sulfate solutions containing tartaric acid. Miyazaki et al. (2012)
32
2 Characterization and Modeling of Lithium Dendrite Growth
also reported suppression of dendrite formation in metallic Zn deposition using zinc oxide electrodes modified with an anion-exchange ionomer (AEI). These improvements are explained by selective ion permeation through the AEI films. These approaches on Zn dendrite suppression and the techniques used in the investigation of Zn dendrite suppression and the technics used in the investigation of Zn dendrite growth are very useful for the investigation and prevention of Li dendrite growth. Although most electrodepositions are a one-time-only process, in a rechargeable Li metal battery, Li metal needs to be plated on or stripped from substrates repeatedly during charge/discharge processes. As a result, Li dendrites will accumulate on the anode and finally lead to many serious problems (see Fig. 1.1) that hamper the practical application of rechargeable Li metal batteries. Therefore, a good understanding of the mechanism of Li dendrite formation and growth is critical to mitigate or further eliminate Li dendrites. Many groups have simulated the Li dendrite formation and growth process, and proposed several meaningful and fundamental models in the last forty years. In order to simplify the simulation conditions, most models were based on a binary electrolyte with a Li salt and polymer; for example, LiClO4 or LiN(SO2CF3)2 (LiTFSI) in polyethylene oxide (PEO). In the open circuit condition, the electrolyte is in a steady state without an ionic concentration gradient; under polarization, the Li+ and anion will diverge and transfer to the negative and positive electrode, respectively. Li+ will obtain an electron and plate on the negative electrode. The speed of Li deposition or consumption of Li+ depends on the applied current density. Although the depletion of Li+ can be macroscopically compensated by the supply of Li+ from the positive electrode, the microscopic ionic distribution near the negative electrode dramatically affects the deposit’s morphology. Therefore, a basic model to simulate Li dendrite starts from the calculation of the concentration gradient in the Li symmetric cell under polarization. Brissot and Chazalviel et al. described the concentration gradient in a cell with a small inter-electrode distance using the following equation (Rosso et al. 2006; Brissot et al. 1999b): @C Jla ðxÞ ¼ @x eDðla þ lLi þ Þ
ð2:1Þ
where J is the effective electrode current density, D is the ambipolar diffusion coefficient, e is the electronic charge, and la and lLi+ are the anionic and Li+ mobilities. From Eq. (2.1), two different conditions for a symmetrical Li/PEO/Li cell can be anticipated, with the inter-electrode distance L and initial Li salt concentration Co: (a) If dC/dx < 2Co/L, the ionic concentration evolves to a steady state where the concentration gradient is constant and varies almost linearly from Ca ¼ Co DCa at the negative electrode to CLi þ ¼ Co þ DCLi þ at the positive electrode, where
2.3 Modeling of Lithium Dendrite Growth
33
DCa DCLi þ
la JL la þ lLi þ eD
ð2:2Þ
(b) If dC/dx > 2Co/L, the ionic concentration goes to zero at the negative electrode at a time called “Sand’s time” s, which varies as eCo 2 s ¼ pD 2Jta ta 1 tLi þ ¼
la la þ lLi þ
ð2:3Þ ð2:4Þ
where ta and tLi+ represent the anionic and Li+ transference number, respectively. Chazalviel indicated that the anionic and Li+ concentrations exhibit different behaviors at the Sand time, leading to an excess of positive charge at the negative electrode. This behavior will result in a local space charge, form a large electric field, and lead to nucleation and growth of Li dendrite. The results of their simulations and experiments confirmed the concentration gradient and the occurrence of dendrites very close to Sand’s time (Brissot et al. 1998, 1999a). Chazalviel (1990) also predicted that the dendrite will grow at the velocity of m ¼ la E
ð2:5Þ
where E is the electric field strength. Monroe and Newman developed the general model describing dendrite growth under galvanostatic conditions applicable to liquid electrolytes (Monroe and Newman 2003). They adopted the Barton and Bockris method (Barton and Bockris 1962) with the addition of thermodynamic reference points and the fact that concentration and potential are not constant during the course of dendrite growth. They calculated the concentration and potential distribution in the cell at different time intervals. It was demonstrated that the dendrite growth rate increases across the electrolyte and depends greatly on the applied current density; this will be discussed in more detail in the following section. Rosso et al. reported a systematic study on the evolution of Li dendrites in LiTFSI-PEO electrolyte involving theoretical calculations as well as experimental data (Rosso et al. 2006). They demonstrated that even though the formation of dendrites has little effect on overall impedance, it significantly decreases interfacial impedance. Based on the impedance data, the value of resistance due to the dendrites could be calculated. In addition, it was observed that dendrites can burn out like a fuse; that is, when the first dendrite reaches the opposite electrode it shorts the circuit and the current density passing through this single dendrite becomes high enough to melt it. Thus, the final short circuit occurs only when the major front of dendrites eventually reaches the opposite electrode.
34
2 Characterization and Modeling of Lithium Dendrite Growth
Although the formation of contiguous and conducting Li dendrites in batteries is often called “dendritic growth,” there are actually several modes of formation and growth: dendrites, whiskers, and “others.” The true dendrite grows from a Li metal surface in a nonaqueous electrolyte by adding material to its tip. The nutrient source is the Li in the electrolyte. Classical models of dendrite growth gave solutions in the form of the tip radius times its velocity, which has the units of diffusivity. Recent electrochemical continuum models (Rosso et al. 2001) and experiments (Bhattacharyya et al. 2010) for Li batteries have found that the dendrite growth is controlled by the tip surface energy, always accelerates across the cell under all conditions, and can be partially mitigated by lowering the limiting current or increasing the cell thickness. The latter two conditions limit the performance of the battery. A second type of growth has been simulated in some Li battery experiments. If the nutrient supply is drawn from the Li metal sheet, growth occurs at the base of a “whisker.” Yamaki (1998) analyzed the stress-assisted whisker growth through cracks in a protective layer (i.e., the separator) on the surface of the Li anode.
2.3.2
Effect of Current Density
It is well known that the effective current density during Li deposition/stripping has a significant impact on the dendrite formation and growth. Generally, low current density results in relatively stable cycling, and conversely, high current density accelerates the degradation process of rechargeable Li metal batteries. The equation of Sand’s time indicates that the dendrite initiation time is proportional to J−2, which indicates that high current density greatly accelerates the formation of Li dendrites. It is worth noting that there are some results showing s * J−1.25 rather than s * J−2 dependence, as reported by (Liu et al. 2010). They attributed this deviation to the local fluctuations of current density. Moreover, the ionic liquid (IL) used in their work acted as a supporting electrolyte. In fact, this is a ternary electrolyte rather than a binary one as assumed by Chazalviel’s model. In the model developed by Monroe and Newman (2003) using liquid electrolyte, tip growth rate (mtip) can be expressed as vtip ¼
Jn V F
ð2:6Þ
where Jn is the effective current density normal to the dendrite (hemispherical) tip, V is the molar volume of Li and F is Faraday’s constant. This equation suggests that the dendrite growth rate is proportional to Jn. Based on Eq. (2.3) and (2.6), the dendrite initiation time could be delayed and the dendrite growth rate could be slowed down if the effective current density could be decreased. By applying a smaller current density, a smoother surface and improved cycling life have been
2.3 Modeling of Lithium Dendrite Growth
35
obtained (Kanamura et al. 1996; Aurbach et al. 2000; Zinigrad et al. 2001; Wang et al. 2000; Crowther and West 2008; Gireaud et al. 2006). According to Chazalviel’s model, an applied current density leads to an ion concentration gradient—high-current density results in near-zero ion concentration at the negative electrode and the formation of Li dendrites at Sand’s time, low current density leads to a minimal and stable ion concentration gradient so no Li dendrites form in this condition. The crossover between the two regimes is the limiting current density J ¼ 2eCo D=ta L
ð2:7Þ
where L is the inter-electrode distance, ta is the anionic transport number. When the current density is low or the inter-electrode distance L is small, there is in principle no Sand behavior and the concentration variation should be small. However, experimental results clearly indicate there are still Li dendrites, just not as serious as those at high-current density. Rosso et al. (2001) and Teyssot et al. (2005) attributed the formation of dendrites to local nonuniformity of the Li/electrolyte interface, which leads to a large concentration variance even in the depleted zone close to the conditions of Chazalviel’s model. Brissot et al. (1998) confirmed this experimentally in a Li| LiTFSI-PEO|Li cell although individual dendrites could deviate from the predicted growth rate. It was demonstrated that at high current densities (when Li deposition becomes diffusion controlled), the onset of dendrite formation matched Sand’s time (zero ion concentration). However, dendrites started to grow earlier with cycling, apparently because of the created defects. At low current densities (i.e., when concentration variations were low), dendrites were also observed in the form of elongated metal filaments (higher aspect ratio), which could be a result of local inhomogeneity. Growth velocities followed Chazalviel’s model (Chazalviel 1990) in both cases. It was later shown by Rosso et al. (2001) that the time of dendrite appearance at low current densities is also proportional to the power of current density even though Sand’s behavior was not expected. It was proposed that the specific properties of LiTFSI-PEO electrolyte cause destabilization of the concentration distribution along the electrode surface. A direct relation between dendritic growth and concentration gradient was clearly demonstrated by employing three independent techniques to measure ion concentrations in the vicinity of dendrites (Brissot et al. 1999a). In addition to the value of current density, the charging styles, galvanostat, or pulse also significantly affects Li dendrite formation and growth. Recent work by Miller and coauthors reported that pulsed charging can effectively suppress Li dendrite formation by as much as 96 %. They proposed a coarse-grained lattice model to explain the mechanism of pulsed charging and revealed that dendrite formation arose from a competition between the time scales of Li+ diffusion and reduction at the anode, with lower overpotential and shorter electrode pulse durations shifting this competition in favor of lower dendrite formation probability (Mayers et al. 2012).
36
2.3.3
2 Characterization and Modeling of Lithium Dendrite Growth
Importance of Interfacial Elastic Strength
Monroe and Newman (2005) further employed linear elasticity theory to develop a kinetic model describing how mechanical properties of polymer electrolytes (shear modulus and Poisson’s ratio) affect roughness on the Li interface. The interface was subjected to a regime of small-amplitude two-dimensional (2D) perturbations. Analytic solutions with specific boundary conditions allowed computing deformation profiles. Compressive stress, deformation stress, and surface tension at the elastic Li interface were calculated as a next step. Incorporation of these parameters into the model gave a prediction of the distribution of exchange current density along the electrode surface. Finally, it was possible to verify that the mechanical properties of the polymer electrolyte stopped amplification of the dendrite growth. It turned out that dendrite suppression can be achieved when the shear modulus of the electrolyte is about twice that of the Li anode (*109 Pa); that is, at least three orders of magnitude higher than that of the studied PEO. As in many other fields, all of the models discussed above have their limitations. For example, the dendrite growth velocity proposed by Monroe and Newman (2003) (Eq. (2.6)) was derived from the growth of a single dendrite without considering the interaction between neighboring dendrites. It was also stated that the Chazalviel theory (Chazalviel 1990) has limited application in real batteries because it applies at currents higher than the limiting current. However, Rosso et al. (2001) and Teyssot et al. (2005) concluded that the Chazalviel model can be extended to low currents due to the nanoscale inhomogeneity in concentration, at least in the case of PEO-based electrolytes. Even though these models still include many simplifications and limitations, they have established a solid foundation for the nucleation and growth mechanisms of dendrites. More importantly, as we will discuss later in this book, several general predications of these models have been used successfully to identify new approaches to suppress dendrite growth, especially during Li metal deposition; for example, using an anode with large surface area to reduce the effective current density, developing a single ion conductor to enhance Li+ transference number, developing an electrolyte with strong shear modulus, and adding supporting electrolyte. These approaches will be discussed in detail in Chap. 3.
References Aaboubi O, Douglade J, Abenaqui X, Boumedmed R, VonHoff J (2011) Influence of tartaric acid on zinc electrodeposition from sulphate bath. Electrochim Acta 56(23):7885–7889 Akolkar R (2013) Mathematical model of the dendritic growth during lithium electrodeposition. J Power Sources 232:23–28. doi:10.1016/j.jpowsour.2013.01.014 Akolkar R (2014) Modeling dendrite growth during lithium electrodeposition at sub-ambient temperature. J Power Sources 246:84–89. doi:10.1016/j.jpowsour.2013.07.056
References
37
Arai J, Okada Y, Sugiyama T, Izuka M, Gotoh K, Takeda K (2015) In situ solid state 7Li NMR observations of lithium metal deposition during overcharge in lithium ion batteries. J Electrochem Soc 162(6):A952–A958. doi:10.1149/2.0411506jes Arakawa M, Tobishima S-I, Nemoto Y, Ichimura M (1993) Lithium electrode cycleability and morphology dependence on current density. J Power Sources 43–44:27–35 Aryanfar A, Brooks D, Merinov BV, Goddard WA III, Colussi AJ, Hoffmann MR (2014) Dynamics of lithium dendrite growth and inhibition: pulse charging experiments and monte carlo calculations. J Phys Chem Lett 5(10):1721–1726. doi:10.1021/jz500207a Aryanfar A, Brooks DJ, Colussi AJ, Merinov BV, Goddard WA III, Hoffmann MR (2015) Thermal relaxation of lithium dendrites. Phys Chem Chem Phys 17(12):8000–8005. doi:10. 1039/c4cp05786d Aurbach D, Cohen Y (1996) The application of atomic force microscopy for the study of Li deposition processes. J Electrochem Soc 143(11):3525–3532 Aurbach D, Cohen Y (1997) Morphological studies of Li deposition processes in LiAsF6/PC solutions by in situ atomic force microscopy. J Electrochem Soc 144(10):3355–3360 Aurbach D, Daroux ML, Faguy PW, Yeager E (1987) Identification of surface films formed on lithium in propylene carbonate solutions. J Electrochem Soc 134(7):1611–1620 Aurbach D, Gofer Y, Langzam J (1989) The correlation between surface chemistry, surface morphology, and cycling efficiency of lithium electrodes in a few polar aprotic systems. J Electrochem Soc 136(11):3198–3205 Aurbach D, Youngman O, Dan P (1990a) The electrochemical behavior of 1,3-Dioxolane-LiClO4 solutions—II. Contaminated solutions. Electrochim Acta 35(3):639–655 Aurbach D, Youngman O, Gofer Y, Meitav A (1990b) The electrochemical behavior of 1,3-Dioxolane-LiClO4 solutions—I. Uncontaminated Solutions. Electrochim Acta 35(3):625– 638 Aurbach D, Daroux M, McDougall G, Yeager EB (1993) Spectroscopic studies of lithium in an ultrahigh vacuum system. J Electroanal Chem 358:63–76. doi:10.1016/0022-0728(93)80431-G Aurbach D, Weissman I, Zaban A, Chusid O (1994) Correlation between surface chemistry, morphology, cycling efficiency and interfacial properties of Li electrodes in solutions containing different Li salts. Electrochim Acta 39:51–71 Aurbach D, Zaban A, Gofer Y, Ely YE, Weissman I, Chusid O, Abramson O (1995) Recent studies of the lithium-liquid electrolyte interface. electrochemical, morphological and spectral studies of a few important systems. J Power Sources 54:76–84 Aurbach D, Zaban A, Ein-Eli Y, Weissman I, Chusid O, Markovsky B, Levi M, Levi E, Schechter A, Granot E (1997) Recent studies on the correlation between surface chemistry, morphology, three-dimensional structures and performance of Li and Li-C intercalation anodes in several important electrolyte systems. J Power Sources 68:91–98 Aurbach D, Weissman I, Yamin H, Elster E (1998) The correlation between charge/discharge rates and morphology, surface chemistry, and performance of Li electrodes and the connection to cycle life of practical batteries. J Electrochem Soc 145(5):1421–1426 Aurbach D, Markovsky B, Levi MD, Levi E, Schechter A, Moshkovich M, Cohen Y (1999) New insights into the interactions between electrode materials and electrolyte solutions for advanced nonaqueous batteries. J Power Sources 81–82:95–111 Aurbach D, Zinigrad E, Teller H, Dan P (2000) Factors which limit the cycle life of rechargeable lithium (metal) batteries. J Electrochem Soc 147:1274–1279 Barton JL, Bockris JOM (1962) The electrolytic growth of dendrites from ionic solutions. Proc R Soc Lond A 268:485–505 Besenhard JO, Gürtler J, Komenda P, Paxinos A (1987) Corrosion protection of secondary lithium electrodes in organic electrolytes. J Power Sources 20:253–258 Bhattacharyya R, Key B, Chen H, Best AS, Hollenkamp AF, Grey CP (2010) In situ NMR observation of the formation of metallic lithium microstructures in lithium batteries. Nat Mater 9(6):504–510. doi:10.1038/nmat2764
38
2 Characterization and Modeling of Lithium Dendrite Growth
Bieker G, Winter M, Bieker P (2015) Electrochemical in situ investigations of SEI and dendrite formation on the lithium metal anode. Phys Chem Chem Phys 17(14):8670–8679. doi:10.1039/ c4cp05865h Brissot C, Rosso M, Chazalviel JN, Baudry P, Lascaud S (1998) In situ study of dendritic growth in lithium/PEO-salt/lithium cells. Electrochim Acta 43(10–11):1569–1574 Brissot C, Rosso M, Chazalviel JN, Lascaud S (1999a) In situ concentration cartography in the neighborhood of dendrites growing in lithium/polymer-electrolyte/lithium cells. J Electrochem Soc 146(12):4393–4400 Brissot C, Rosso M, Chazalviel JN, Lascaud S (1999b) Dendritic growth mechanisms in lithium/polymer cells. J Power Sources 81–82:925–929. doi:10.1016/S0378-7753(98)00242-0 Chandrashekar S, Trease NM, Chang HJ, Du L-S, Grey CP, Jerschow A (2012) 7Li MRI of Li batteries reveals location of microstructural lithium. Nat Mater 11(4):311–315. doi:http://www. nature.com/nmat/journal/v11/n4/abs/nmat3246.html#supplementary-information Chang HJ, Trease NM, Ilott AJ, Zeng D, Du L-S, Jerschow A, Grey CP (2015) Investigating Li microstructure formation on Li anodes for lithium batteries by in situ 6Li/7Li NMR and SEM. J Phys Chem C 119(29):16443–16451. doi:10.1021/acs.jpcc.5b03396 Chazalviel JN (1990) Electrochemical aspects of the generation of ramified metallic electrodeposits. Phys Rev A 42(12):7355–7367. doi:10.1103/PhysRevA.42.7355 Chen L, Zhang HW, Liang LY, Liu Z, Qi Y, Lu P, Chen J, Chen L-Q (2015) Modulation of dendritic patterns during electrodeposition: a nonlinear phase-field model. J Power Sources 300:376–385. doi:10.1016/j.jpowsour.2015.09.055 Cheng X-B, Zhang R, Zhao C-Z, Wei F, Zhang J-G, Zhang Q (2015) A review of solid electrolyte interphases on lithium metal anode. Adv Sci: 1500213 Choi N-S, Koo B, Yeon J-T, Lee KT, Kim D-W (2011) Effect of a novel amphipathic ionic liquid on lithium deposition in gel polymer electrolytes. Electrochim Acta 56(21):7249–7255. doi:10. 1016/j.electacta.2011.06.058 Cogswell DA (2015) Quantitative phase-field modeling of dendritic electrodeposition. Phys Rev E 92(1):011301. doi:10.1103/PhysRevE.92.011301 Cohen YS, Cohen Y, Aurbach D (2000) Micromorphological studies of lithium electrodes in alkyl carbonate solutions using in situ atomic force microscopy. J Phys Chem B 104:12282–12291 Crowther O, West AC (2008) Effect of electrolyte composition on lithium dendrite growth. J Electrochem Soc 155(11):A806–A811. doi:10.1149/1.2969424 Dampier FW, Brummer SB (1977) The cycling behavior of the lithium electrode in LiAsF6/methyl acetate solutions. Electrochim Acta 22:1339–1345 Deng X, Hu MY, Wei X, Wang W, Chen Z, Liu J, Hu JZ (2015) Natural abundance 17O nuclear magnetic resonance and computational modeling studies of lithium based liquid. J Power Sources 285:146–155. doi:10.1016/j.jpowsour.2015.03.091 Dey AN (1976) S.E.M. studies of the Li-film growth and the voltage-delay phenomenon associated with the lithium-thionyl chloride inorganic electrolyte system. Electrochim Acta 21 (5):377–382 Diggle JW, Despic AR, Bockris JOM (1969) The mechanism of the dendritic electrocrystallization of zinc. J Electrochem Soc 112:1503–1514 Ding F, Xu W, Graff GL, Zhang J, Sushko ML, Chen X, Shao Y, Engelhard MH, Nie Z, Xiao J, Liu X, Sushko PV, Liu J, Zhang J-G (2013) Dendrite-free lithium deposition via self-healing electrostatic shield mechanism. J Am Chem Soc 135(11):4450–4456. doi:10.1021/ja312241y Ding F, Xu W, Chen X, Zhang J, Shao Y, Engelhard MH, Zhang Y, Blake TA, Graff GL, Liu X, Zhang J-G (2014) Effects of cesium cations in lithium deposition via self-healing electrostatic shield mechanism. J Phys Chem C 118(8):4043–4049. doi:10.1021/jp4127754 Dollé M, Sannier L, Beaudoin B, Trentin M, Tarascon J-M (2002) Live scanning electron microscope observations of dendritic growth in lithium/polymer cells. Electrochem Solid-State Lett 5(12):A286–A289. doi:10.1149/1.1519970 Fujieda T, Yamamoto N, Saito K, Ishibashi T, Honjo M, Koike S, Wakabayashi N, Higuchi S (1994) Surface of lithium electrodes prepared in Ar + CO2 Gas. J Power Sources 52:197–200. doi:10.1016/0378-7753(94)01961-4
References
39
Ghassemi H, Au M, Chen N, Heiden PA, Yassar RS (2011) Real-time observation of lithium fibers growth inside a nanoscale lithium-ion battery. Appl Phys Lett 99(12):123113. doi:10.1063/1. 3643035 Gireaud L, Grugeon S, Laruelle S, Yrieix B, Tarascon JM (2006) Lithium metal stripping/plating mechanisms studies: a metallurgical approach. Electrochem Commun 8(10):1639–1649. doi:10.1016/j.elecom.2006.07.037 Gu M, Parent LR, Mehdi BL, Unocic RR, McDowell MT, Sacci RL, Xu W, Connell JG, Xu P, Abellan P, Chen X, Zhang Y, Perea DE, Evans JE, Lauhon LJ, Zhang JG, Liu J, Browning ND, Cui Y, Arslan I, Wang CM (2013) Demonstration of an electrochemical liquid cell for operando transmission electron microscopy observation of the lithiation/delithiation behavior of Si nanowire battery anodes. Nano Lett 13(12):6106–6112. doi:10.1021/nl403402q Harry KJ, Hallinan DT, Parkinson DY, MacDowell AA, Balsara NP (2014) Detection of subsurface structures underneath dendrites formed on cycled lithium metal electrodes. Nat Mater 13(1):69–73. doi:10.1038/nmat3793 Harry KJ, Liao X, Parkinson DY, Minor aM, Balsaraa NP (2015) Electrochemical Deposition and Stripping Behavior of Lithium Metal across a Rigid Block Copolymer Electrolyte Membrane. J Electrochem Soc 162(14):A2699–A2706 Hernandez-Maya R, Rosas O, Saunders J, Castaneda H (2015) Dynamic characterization of dendrite deposition and growth in Li-surface by electrochemical impedance spectroscopy. J Electrochem Soc 162(4):A687–A696. doi:10.1149/2.0561504jes Howlett PC, MacFarlane DR, Hollenkamp AF (2003) A sealed optical cell for the study of lithium-electrode|electrolyte interfaces. J Power Sources 114(2):277–284. doi:10.1016/s03787753(02)00603-1 Huang JY, Zhong L, Wang CM, Sullivan JP, Xu W, Zhang LQ, Mao SX, Hudak NS, Liu XH, Subramanian A, Fan H, Qi L, Kushima A, Li J (2010) In situ observation of the electrochemical lithiation of a single SnO2 nanowire electrode. Science 330:1515–1520 Ishikawa M, Morita M, Matsuda Y (1997) In situ scanning vibrating electrode technique for lithium metal anodes. J Power Sources 68:501–505 Jäckle M, Groß A (2014) Microscopic properties of lithium, sodium, and magnesium battery anode materials related to possible dendrite growth. J Chem Phys 141(17):174710. doi:10. 1063/1.4901055 Kanamura K, Tamura H, Shiraishi S, Takehara Z-I (1995a) Morphology and chemical compositions of surface films of lithium deposited on a Ni substrate in nonaqueous electrolytes. J Electroanal Chem 394:49–62 Kanamura K, Tamura H, Shiraishi S, Takehara Z-I (1995b) XPS analysis of lithium surfaces following immersion in various solvents containing LiBF4. J Electrochem Soc 142(2):340–347 Kanamura K, Shiraishi S, Takeharo Z-I (1996) Electrochemical deposition of very smooth lithium using nonaqueous electrolytes containing HF. J Electrochem Soc 143(7):2187–2197 Kanamura K, Takezawa H, Shiraishi S, Takehara Z-I (1997) Chemical reaction of lithium surface during immersion in LiClO4 or LiPF6/DEC electrolyte. J Electrochem Soc 144(6):1900–1906 Kominato A, Yasukawa E, Sato N, Ijuuin T, Asahina H, Mori S (1997) Analysis of surface films on lithium in various organic electrolytes. J Power Sources 68:471–475 Lan CJ, Lee CY, Chin TS (2007) Tetra-alkyl ammonium hydroxides as inhibitors of Zn dendrite in Zn-based secondary batteries. Electrochim Acta 52(17):5407–54116 Lee Y-G, Kyhm K, Choi N-S, Ryu KS (2006) Submicroporous/microporous and compatible/incompatible multi-functional dual-layer polymer electrolytes and their interfacial characteristics with lithium metal anode. J Power Sources 163(1):264–268. doi:10.1016/j. jpowsour.2006.05.008 Lee YM, Seo JE, Lee Y-G, Lee SH, Cho KY, Park J-K (2007) Effects of triacetoxyvinylsilane as SEI layer additive on electrochemical performance of lithium metal secondary battery. Electrochem Solid-State Lett 10(9):A216–A219. doi:10.1149/1.2750439 Li W, Zheng H, Chu G, Luo F, Zheng J, Xiao D, Li X, Gu L, Li H, Wei X, Chen Q, Chen L (2014a) Effect of electrochemical dissolution and deposition order on lithium dendrite formation: a top view investigation. Faraday Discuss 176:109–124. doi:10.1039/c4fd00124a
40
2 Characterization and Modeling of Lithium Dendrite Growth
Li Z, Huang J, Liaw BY, Metzler V, Zhang J (2014b) A review of lithium deposition in lithium-ion and lithium metal secondary batteries. J Power Sources 254:168–182. doi:10.1016/ j.jpowsour.2013.12.099 Liu S, Imanishi N, Zhang T, Hirano A, Takeda Y, Yamamoto O, Yang J (2010) Lithium dendrite formation in Li/Poly(ethylene oxide)–lithium Bis(trifluoromethanesulfonyl)imide and N-Methyl-N-propylpiperidinium Bis(trifluoromethanesulfonyl)imide/Li cells. J Electrochem Soc 157(10):A1092–A1098. doi:10.1149/1.3473790 Liu XH, Zhong L, Zhang LQ, Kushima A, Mao SX, Li J, Ye ZZ, Sullivan JP, Huang JY (2011) Lithium fiber growth on the anode in a nanowire lithium ion battery during charging. Appl Phys Lett 98(18):183107. doi:10.1063/1.3585655 López CM, Vaughey JT, Dees DW (2009) Morphological transitions on lithium metal anodes. J Electrochem Soc 156(9):A726–A729. doi:10.1149/1.3158548 López CM, Vaughey JT, Dees DW (2012) Insights into the role of interphasial morphology on the electrochemical performance of lithium electrodes. J Electrochem Soc 159(6):A873–A886. doi:10.1149/2.100206jes Lv D, Shao Y, Lozano T, Bennett WD, Graff GL, Polzin B, Zhang J-G, Engelhard MH, Saenz NT, Henderson WA, Bhattacharya P, Liu J, Xiao J (2015) Failure mechanism for fast-charged lithium metal batteries with liquid electrolytes. Adv Energy Mater 5(3):1400993. doi:10.1002/ aenm.201400993 Magan RV, Sureshkumar R, Lin B (2003) Influence of surface reaction rate on the size dispersion of interfacial nanostructures. J Phys Chem B 107:10513–10520 Matsuda Y, Ishikawa M, Yoshitake S, Morita M (1995) Characterization of the lithium-organic electrolyte interface containing inorganic and organic additives by in situ techniques. J Power Sources 54:301–305 Mayers MZ, Kaminski JW, Miller TF (2012) Suppression of dendrite formation via pulse charging in rechargeable lithium metal batteries. J Phys Chem C 116(50):26214–26221. doi:10.1021/ jp309321w Mehdi BL, Gu M, Parent LR, Xu W, Nasybulin EN, Chen X, Unocic RR, Xu P, Welch DA, Abellan P, Zhang JG, Liu J, Wang CM, Arslan I, Evans J, Browning ND (2014) In-situ electrochemical transmission electron microscopy for battery research. Microsc Microanal 20 (2):484–492. doi:10.1017/S1431927614000488 Mehdi BL, Qian J, Nasybulin E, Park C, Welch DA, Faller R, Mehta H, Henderson WA, Xu W, Wang CM, Evans JE, Liu J, Zhang JG, Mueller KT, Browning ND (2015) Observation and quantification of nanoscale processes in lithium batteries by operando electrochemical (S)TEM. Nano Lett 15(3):2168–2173. doi:10.1021/acs.nanolett.5b00175 Mendoza-Huízar LH, Rios-Reyes CH, Gómez-Villegas MG (2009) Zinc electrodeposition from chloride solutions onto glassy carbon electrode. J Mex Chem Soc 53(4):243–247 Miyazaki K, Lee YS, Fukutsuka T, Abe T (2012) Suppression of dendrite formation of zinc electrodes by the modification of anion-exchange ionomer. Electrochemistry 80(10):725–727. doi:10.5796/electrochemistry.80.725 Mogi R, Inaba M, Iriyama Y, Abe T, Ogumi Z (2002a) In situ atomic force microscopy study on lithium deposition on nickel substrates at elevated temperatures. J Electrochem Soc 149(4): A385–A390. doi:10.1149/1.1454138 Mogi R, Inaba M, Iriyama Y, Abe T, Ogumi Z (2002b) Surface film formation on nickel electrodes in a propylene carbonate solution at elevated temperatures. J Power Sources 108:163–173 Mogi R, Inaba M, Jeong S-K, Iriyama Y, Abe T, Ogumi Z (2002c) Effects of some organic additives on lithium deposition in propylene carbonate. J Electrochem Soc 149(12):A1578– A1583. doi:10.1149/1.1516770 Monroe C, Newman J (2003) Dendrite growth in lithium/polymer systems: a propagation model for liquid electrolytes under galvanostatic conditions. J Electrochem Soc 150(10):A1377– A1384. doi:10.1149/1.1606686 Monroe C, Newman J (2004) The effect of interfacial deformation on electrodeposition kinetics. J Electrochem Soc 151(6):A880–A886. doi:10.1149/1.1710893
References
41
Monroe C, Newman J (2005) The impact of elastic deformation on deposition kinetics at lithium/polymer interfaces. J Electrochem Soc 152(2):A396–A404. doi:10.1149/1.1850854 Morigaki K-I (2002) Analysis of the interface between lithium and organic electrolyte solution. J Power Sources 104:13–23 Morigaki K-I, Ohta A (1998) Analysis of the surface of lithium in organic electrolyte by atomic force microscopy, fourier transform infrared spectroscopy and scanning auger electron microscopy. J Power Sources 76:159–166 Nagao M, Hayashi A, Tatsumisago M, Kanetsuku T, Tsuda T, Kuwabata S (2013) In situ SEM study of a lithium deposition and dissolution mechanism in a bulk-type solid-state cell with a Li2S-P2S5 solid electrolyte. Phys Chem Chem Phys 15(42):18600–18606. doi:10.1039/ c3cp51059j Naoi K, Mori M, Shinagawa Y (1996) Study of deposition and dissolution processes of lithium in carbonate-based solutions by means of the quartz-crystal microbalance. J Electrochem Soc 143 (8):2517–2522 Naoi K, Mori M, Naruoka Y, Lamanna WM, Atanasoski R (1999) The surface film formed on a lithium metal electrode in a new imide electrolyte, lithium Bis(perfluoroethylsulfonylimide) [LiN(C2F5SO2)2]. J Electrochem Soc 146(2):462–469 Naudin C, Bruneel JL, Chami M, Desbat B, Grondin J, Lassègues JC, Servant L (2003) Characterization of the lithium surface by infrared and raman spectroscopies. J Power Sources 124(2):518–525. doi:10.1016/s0378-7753(03)00798-5 Nazri G, Muller RH (1985a) In situ X-Ray diffraction of surface layers on lithium in nonaqueous electrolyte. J Electrochem Soc 132(6):1385–1387 Nazri G, Muller RH (1985b) Composition of surface layers on Li electrodes in PC, LiClO4 of very low water content. J Electrochem Soc 132(9):2050–2054 Nazri G, Muller RH (1985c) Effect of residual water in propylene carbonate on films formed on lithium. J Electrochem Soc 132(9):2054–2058 Neudecker BJ, Dudney NJ, Bates JB (2000) “Lithium-Free” thin-film battery with in situ plated Li anode. J Electrochem Soc 147(2):517–523 Nishida T, Nishikawa K, Rosso M, Fukunaka Y (2013) Optical observation of Li dendrite growth in ionic liquid. Electrochim Acta 100:333–341. doi:10.1016/j.electacta.2012.12.131 Nishikawa K, Fukunaka Y, Sakka T, Ogata YH, Selman JR (2007) Measurement of concentration profiles during electrodeposition of Li metal from LiPF6-PC electrolyte solution. J Electrochem Soc 154(10):A943–A948. doi:10.1149/1.2767404 Nishikawa K, Mori T, Nishida T, Fukunaka Y, Rosso M, Homma T (2010) In situ observation of dendrite growth of electrodeposited Li metal. J Electrochem Soc 157(11):A1212–A1217. doi:10.1149/1.3486468 Nishikawa K, Mori T, Nishida T, Fukunaka Y, Rosso M (2011) Li dendrite growth and Li+ ionic mass transfer phenomenon. J Electroanal Chem 661(1):84–89. doi:10.1016/j.jelechem.2011. 06.035 Nishikawa K, Naito H, Kawase M, Nishida T (2012) Morphological variation of electrodeposited Li in ionic liquid. ECS Trans 41:3–10 Odziemkowski M, Krell M, Irish DE (1992) A raman microprobe in situ and ex situ study of film formation at lithium/organic electrolyte interfaces. J Electrochem Soc 139(11):3052–3063 Orsini F, Du Pasquier A, Beaudoin B, Tarascon JM, Trentin M, Langenhuizen N, De Beer E, Notten P (1998) In situ scanning electron microscopy (SEM) observation of interfaces within plastic lithium batteries. J Power Sources 76:19–29 Orsini F, du Pasquier A, Beaudouin B, Tarascon JM, Trentin M, Langenhuizen N, de Beer E, Notten P (1999) In situ SEM study of the interfaces in plastic lithium cells. J Power Sources 81–82:918–921. doi:10.1016/S0378-7753(98)00241-9 Osaka T, Homma T, Momma T, Yarimizu H (1997a) In situ observation of lithium deposition processes in solid polymer and gel electrolytes. J Electroanal Chem 421:153–156 Osaka T, Momma T, Matsumoto Y, Uchida Y (1997b) Surface characterization of electrodeposited lithium anode with enhanced cycleability obtained by CO2 addition. J Electrochem Soc 144(5):1709–1713
42
2 Characterization and Modeling of Lithium Dendrite Growth
Ota H, Wang X, Yasukawa E (2004a) Characterization of lithium electrode in lithium imides/ethylene carbonate, and cyclic ether electrolytes. I. Surface morphology and lithium cycling efficiency. J Electrochem Soc 151(3):A427–A436. doi:10.1149/1.1644136 Ota H, Shima K, Ue M, Yamaki J-I (2004b) Effect of vinylene carbonate as additive to electrolyte for lithium metal anode. Electrochim Acta 49(4):565–572. doi:10.1016/j.electacta.2003.09.010 Ota H, Sakata Y, Wang X, Sasahara J, Yasukawa E (2004c) Characterization of lithium electrode in lithium imides/ethylene carbonate and cyclic ether electrolytes. II. Surface chemistry. J Electrochem Soc 151(3):A437–A446. doi:10.1149/1.1644137 Qian J, Xu W, Bhattacharya P, Engelhard M, Henderson WA, Zhang Y, Zhang J-G (2015a) Dendrite-free Li deposition using trace-amounts of water as an electrolyte additive. Nano Energy 15:135–144. doi:10.1016/j.nanoen.2015.04.009 Qian J, Henderson WA, Xu W, Bhattacharya P, Engelhard M, Borodin O, Zhang JG (2015b) High rate and stable cycling of lithium metal anode. Nat Commun 6:6362. doi:10.1038/ ncomms7362 Rey I, Lassègues JC, Baudry P, Majastre H (1998a) Study of a lithium battery by confocal Raman microspectrometry. Electrochim Acta 43(10–11):1539–1544 Rey I, Bruneel J-L, Grondin J, Servant L, Lassègues J-C (1998b) Raman spectroelectrochemistry of a lithium/polymer electrolyte symmetric cell. J Electrochem Soc 145(9):3034–3042 Rosso M, Gobron T, Brissot C, Chazalviel JN, Lascaud S (2001) Onset of dendritic growth in lithium/polymer cells. J Power Sources 97–98:804–806 Rosso M, Brissot C, Teyssot A, Dollé M, Sannier L, Tarascon J-M, Bouchet R, Lascaud S (2006) Dendrite short-circuit and fuse effect on Li/polymer/Li cells. Electrochim Acta 51(25):5334– 5340. doi:10.1016/j.electacta.2006.02.004 Ryou M-H, Lee DJ, Lee J-N, Lee YM, Park J-K, Choi JW (2012) Excellent cycle life of lithium-metal anodes in lithium-ion batteries with mussel-inspired polydopamine-coated separators. Adv Energy Mater 2(6):645–650. doi:10.1002/aenm.201100687 Sacci RL, Dudney NJ, More KL, Parent LR, Arslan I, Browning ND, Unocic RR (2014) Direct visualization of initial SEI morphology and growth kinetics during lithium deposition by in situ electrochemical transmission electron microscopy. Chem Commun 50(17):2104–2107. doi:10. 1039/c3cc49029g Sagane F, Ikeda K-I, Okita K, Sano H, Sakaebe H, Iriyama Y (2013) Effects of current densities on the lithium plating morphology at a lithium phosphorus oxynitride glass electrolyte/copper thin film interface. J Power Sources 233:34–42. doi:10.1016/j.jpowsour.2013.01.051 Sano H, Sakaebe H, Matsumoto H (2011) Observation of electrodeposited lithium by optical microscope in room temperature ionic liquid-based electrolyte. J Power Sources 196(16):6663– 6669. doi:10.1016/j.jpowsour.2010.12.023 Sawada Y, Dougherty A, Gollub JP (1986) Dendritic and fractal patterns in electrolytic metal deposits. Phys Rev Lett 56(12):1260–1263. doi:10.1103/PhysRevLett.56.1260 Selim R, Bro P (1974) Some observations on rechargeable lithium electrodes in a propylene carbonate electrolyte. J Electrochem Soc 121(11):1457–1459 Shiraishi S, Kanamura K (1998) The observation of electrochemical dissolution of lithium metal using electrochemical quartz crystal microbalance and in-situ tapping mode atomic force microscopy. Langmuir 14:7082–7086 Shiraishi S, Kanamura K, Zi Takehara (1999) Surface condition changes in lithium metal deposited in nonaqueous electrolyte containing HF by dissolution-deposition cycles. J Electrochem Soc 146(5):1633–1639 Shiraishi S, Kanamura K, Takehara Z-I (2001) Imaging for uniformity of lithium metal surface using tapping mode-atomic force and surface potential microscopy. J Phys Chem B 105:123– 134 Stark JK, Ding Y, Kohl PA (2011) Dendrite-free electrodeposition and reoxidation of lithium-sodium alloy for metal-anode battery. J Electrochem Soc 158(10):A1100–A1105. doi:10.1149/1.3622348
References
43
Stark JK, Ding Y, Kohl PA (2013) Nucleation of electrodeposited lithium metal: dendritic growth and the effect of co-deposited sodium. J Electrochem Soc 160(9):D337–D342. doi:10.1149/2. 028309jes Steiger J (2015) Mechanisms of dendrite growth in lithium metal batteries. PhD Thesis, Karlsruhe Institute of Technology (KIT), Germany Steiger J, Kramer D, Mönig R (2014a) Mechanisms of dendritic growth investigated by in situ light microscopy during electrodeposition and dissolution of lithium. J Power Sources 261:112–119. doi:10.1016/j.jpowsour.2014.03.029 Steiger J, Kramer D, Mönig R (2014b) Microscopic observations of the formation, growth and shrinkage of lithium moss during electrodeposition and dissolution. Electrochim Acta 136:529–536. doi:10.1016/j.electacta.2014.05.120 Steiger J, Richter G, Wenk M, Kramer D, Mönig R (2015) Comparison of the growth of lithium filaments and dendrites under different conditions. Electrochem Commun 50:11–14. doi:10. 1016/j.elecom.2014.11.002 Tang M, Albertus P, Newman J (2009) Two-dimensional modeling of lithium deposition during cell charging. J Electrochem Soc 156(5):A390–A399. doi:10.1149/1.3095513 Teyssot A, Belhomme C, Bouchet R, Rosso M, Lascaud S, Armand M (2005) Inter-electrode in situ concentration cartography in lithium/polymer electrolyte/lithium cells. J Electroanal Chem 584(1):70–74. doi:10.1016/j.jelechem.2005.01.037 Thompson RS, Schroeder DJ, López CM, Neuhold S, Vaughey JT (2011) Stabilization of lithium metal anodes using silane-based coatings. Electrochem Commun 13(12):1369–1372. doi:10. 1016/j.elecom.2011.08.012 Voss RF, Tomkiewicz M (1985) Computer simulation of dendritic electrodeposition. J Electrochem Soc 132(2):371–375 Wang X, Yasukawa E, Kasuya S (2000) Lithium imide electrolytes with two-oxygen-atom-containing cycloalkane solvents for 4 V lithium metal rechargeable batteries. J Electrochem Soc 147(7):2421–2426 Wenzel S, Leichtweiss T, Krüger D, Sann J, Janek J (2015) Interphase formation on lithium solid electrolytes—an in situ approach to study interfacial reactions by photoelectron spectroscopy. Solid State Ionics 278:98–105. doi:10.1016/j.ssi.2015.06.001 Yamaki J-I, Tobishima S-I, Hayashi K, Saito K, Nemoto Y, Arakawa M (1998) A consideration of the morphology of electrochemically deposited lithium in an organic electrolyte. J Power Sources 74:219–227 Yang L, Smith C, Patrissi C, Schumacher CR, Lucht BL (2008) Surface reactions and performance of non-aqueous electrolytes with lithium metal anodes. J Power Sources 185(2):1359–1366. doi:10.1016/j.jpowsour.2008.09.037 Yoon S, Lee J, Kim S-O, Sohn H-J (2008) Enhanced cyclability and surface characteristics of lithium batteries by Li–Mg co-deposition and addition of HF acid in electrolyte. Electrochim Acta 53(5):2501–2506. doi:10.1016/j.electacta.2007.10.019 Yoshimatsu I, Hirai T, Yamaki J-I (1988) Lithium electrode morphology during cycling in lithium cells. J Electrochem Soc 135(10):2422–2427 Zhamu A, Chen G, Liu C, Neff D, Fang Q, Yu Z, Xiong W, Wang Y, Wang X, Jang BZ (2012) Reviving rechargeable lithium metal batteries: enabling next-generation high-energy and high-power cells. Energy Environ Sci 5(2):5701–5707. doi:10.1039/c2ee02911a Zhang Y, Qian J, Xu W, Russell SM, Chen X, Nasybulin E, Bhattacharya P, Engelhard MH, Mei D, Cao R, Ding F, Cresce AV, Xu K, Zhang JG (2014) Dendrite-free lithium deposition with self-aligned nanorod structure. Nano Lett 14(12):6889–6896. doi:10.1021/nl5039117 Zinigrad E, Aurbach D, Dan P (2001) Simulation of galvanostatic growth of polycrystalline Li deposits in rechargeable Li batteries. Electrochim Acta 46:1863–1869
Chapter 3
High Coulombic Efficiency of Lithium Plating/Stripping and Lithium Dendrite Prevention
As described in Chaps. 1 and 2, the Li plating morphology and the CE—both of which are critical for the safety characteristics and cyclability of a Li metal battery —are intricately linked with one another. Almost all the factors that lead to significant dendritic growth also lead to a lower CE and vice versa. This is because the needle-like (“dendritic”) growth results in a greater amount of Li metal surface exposure to the electrolyte. This freshly formed Li metal surface is thermodynamically unstable against the electrolyte solvent molecules, ions, additives, and impurities with which it quickly reacts to form degradation products which may or may not fully passivate the Li surface from further reactions. This continuously consumes both the electrolyte and Li and thus leads to a lower CE. To have a high CE, the side reactions must be minimized. The proclivity for these to occur is proportional to the chemical and electrochemical activity of native Li with the neighboring electrolyte, as well as to the surface area of the deposited Li. Therefore, a high Li cycling CE may be achieved by the use of less reactive electrolyte components, a reduction in the exposed Li surface and the rapid passivation of this surface. The strong correlation between the Li plating/striping (or Li cycling) CE and Li morphology suggests that these should be discussed in tandem in Chap. 3 when considering the impact of differing Li cycling environments and conditions. Special cases of dendrite-free Li plating of highly aligned nanorod structures, but with a low CE (related to the high surface area of the Li nanorods) will also be discussed (Qian et al. 2015a; Ding et al. 2013b).
3.1
Coulombic Efficiency of Lithium Plating/Stripping
The scientific literature associated with Li plating/stripping frequently discusses the efficiency of this process, but the methods used to determine the CE vary and often result in significantly different values. The CE is often variable with cycling © Springer International Publishing Switzerland 2017 J.-G. Zhang et al., Lithium Metal Anodes and Rechargeable Lithium Metal Batteries, Springer Series in Materials Science 249, DOI 10.1007/978-3-319-44054-5_3
45
3 High Coulombic Efficiency of Lithium Plating/Stripping …
46
conditions (current density, temperature) and may change considerably upon extended cycling or aging. Thus, reporting a specific value of the CE for a given electrolyte can be highly misleading without providing details of testing conditions. Method 1 Using this method, the Li cycling CE is defined by Aurbach (1999):
CE ¼
QS QD
ð3:1Þ
where QS and QD are the charges for Li stripping (dissolution) and plating (deposition) determined for each cycle, respectively. This method is often used when depositing Li on inert electrodes such as nickel (Ni), Cu, and tungsten (W). Method 2 This method applies an initial known amount of deposition charge (QT) to plate Li metal on the working electrode (often Ni or Cu). Alternatively, if Li metal is used as the electrode, QT is determined from the electrode thickness and density. A fraction of this charge, i.e., the cycling charge (QD), is cycled across the cell for the stripping and subsequent plating with the overpotential for the process monitored during the cycling. When the overvoltage significantly increases (indicating depletion of all of the deposited Li or an increased impedance due to a highly resistive interphase), the test is completed. The CE from this method can therefore be determined by (Aurbach 1999) QT CEavg ¼ 1 100 % NQD
ð3:2Þ
where N is the total number of cycles for the cell (Appetecchi et al. 1998, 1999). In practice, it may take a long time to exhaust all Li (QT) deposited in the initial cycle. Therefore, a simplified formula for average CE in n cycles (n < N) has been used: h CEnavg ¼
QD
CEnavg QT Qr n
QD
i ¼
nQD þ Qr 100 % nQD þ QT
ð3:3Þ
where Qr is the maximum amount of Li stripped from the working electrode after n cycles (Aurbach et al. 1989; Ding et al. 2013b). In the extreme case of n = N, Eq. (3.3) is equivalent to Eq. (3.2). For symmetric Li||Li cells, Eq. (3.2) becomes (Appetecchi and Passerini 2002): QT CEavg ¼ 1 100 % 2NQD
ð3:4Þ
3.1 Coulombic Efficiency of Lithium Plating/Stripping
47
The factor of two accounts for the two steps in each cycle in which the Li is reduced at one electrode and oxidized at the other. This effectively doubles the losses with respect to a battery in which the Li is only oxidized during the discharge and reduced during the charge (Appetecchi and Passerini 2002). This equation is derived from the single-step-efficiency equation: CE ¼
ðQD QL Þ 100 % QD
ð3:5Þ
where QL is the charge for the lost Li in each step (if QL = 0, the CE = 100 %). With Li electrode, however, QL can only be determined by cycling the cell until all of the charge for the Li is lost, i.e., QT = 2nQL for a Li||Li cell (at which point the overpotential increases rapidly). By substituting this into Eq. (3.5), the CEavg Eq. (3.4) can be derived. Method 3 Some reports use a figure-of-merit (FOM) instead of CE (Abraham and Goldman 1983) with FOMLi ¼
nQD 1 ¼ 1 CEavg QT
ð3:6Þ
The FOMLi is thus [the total accumulated cell capacity]/[theoretical Li capacity or Li turnover number]. This value facilitates the recognition of differences in Li cyclability since an increase in the CE from 0.98 to 0.99 corresponds to an increase in the FOM from 50 to 100 which signifies a doubling of the cell cycle life (Arakawa et al. 1999).
3.2
Electrolyte and In Situ Formed Solid Electrolyte Interphase
The components of electrolytes used for Li+ cation transport vary widely (Marcinek et al. 2015). Although ionic conductivity is most often used as a prime factor for electrolyte optimization since it is one of the determinants for the overall cell resistance and thus the power capability of a cell (Dudley et al. 1991), this property is of little utility for determining the stability of a given electrolyte with Li metal (Tobishima et al. 1995). The consequence of this is that the Li plating characteristics (i.e., morphology, reactivity, etc.) and their relationship to the plating/stripping current density are generally not directly correlated with variations in the electrolyte conductivity. Conditions that favor a high cycling CE include the following:
48
3 High Coulombic Efficiency of Lithium Plating/Stripping …
1. uniform (compact) Li deposition which minimizes the surface area of the plated Li exposed to the electrolyte, 2. the formation of a stable passivation layer which protects the Li from further reactions with electrolyte components—if such a layer is elastic, then it can stretch and contract as necessary to conform to changes in the Li volume and may aid in the redeposition of Li underneath such a layer thus reducing the exposure of fresh Li to the electrolyte (Aurbach 1999). In particular, the deposition morphology is a function of the electrolyte solvent (s), salt(s), salt concentration, current density, temperature, and other factors. The SEI layer formed is often strongly influenced by contaminants or additives, even at the ppm level. Comparisons between the various studies reported in the literature must take this into consideration. In addition, in contrast with graphite anodes in which the passivating layer is supported on the graphite surface, the surface layers that form on electrodeposited Li may be, and usually are, disrupted when the Li is stripped away during cell discharge (for nodular deposits) or lack electrical contact with the current source (for needle-like deposits, which generate dead Li) (Fig. 2.17). Although passivation films are known to form on the Li anode surface in batteries such as Li||SOCl2 primary cells, the role of this layer for Li stabilization was emphasized by Dey (1977) and (Peled 1979, 1983; Peled et al. 1997) who dubbed it a “solid electrolyte interphase” (SEI). The contributing components to the SEI and the effect that this has on Li plating/stripping will be discussed below. Note, however, that the study of the composition of the SEI layer is complicated by many factors, including the continuous evolution of the SEI during cycling/storage, changes that may occur from rinsing the surface (prior to ex situ characterization) or due to high vacuum conditions and transformations of the components due to Ar+ sputtering during XPS depth analysis (Verma et al. 2010; Edström et al. 2006; Ota et al. 2004c). Thus, some caution should be exercised when interpreting the results reported in the diverse studies available in the literature. Typically, as-received Li metal has surface coatings consisting principally of Li2O (in contact with the Li) and an outer coating of LiOH and Li2CO3 (Hong et al. 2004; Kanamura et al. 1992, 1994b, 1995b, d, 1997; Shiraishi et al. 1995). Once the Li contacts the electrolyte, these native layers are transformed by both chemical and electrochemical reactions with the electrolyte components. The SEI films formed on Li in contact with an electrolyte are typically multi-layered with the species becoming more inorganic in nature (i.e., more fully reduced) the closer they are to the Li surface (Schechter et al. 1999). Over time, the interphasial layer may continue to grow in thickness due to continuous reactions, but the SEI composition may also be continuously transformed (Aurbach et al. 1994a, 1995b). Impurities such as dissolved gases (e.g., O2, CO2), H2O, or others are often the most reactive constituents of the electrolyte and, as such, these reactions may dominate the characteristics of the as-formed SEI layers. For solvents that are relatively reactive (e.g., esters and carbonates), solvent degradation is the next most important factor
3.2 Electrolyte and In Situ Formed Solid Electrolyte Interphase
49
in determining the composition of the SEI. If the electrolytes instead have solvents that are more stable to reduction (e.g., ethers), then anion degradation may be very influential in the SEI’s composition (Aurbach et al. 1994b).
3.2.1
Influence of Solvents
Molecular orbital (MO) theory allows for an approximate estimation of the oxidative and reductive stability of a molecule or ion. A higher level of the highest occupied molecular orbital (HOMO) energy indicates that the species is a stronger electron donor and thus more susceptible to oxidation, while a lower level of the lowest unoccupied molecular orbital (LUMO) energy indicates a stronger electron acceptor, which is more readily reduced. The theory allows for an approximate estimation of the oxidative and reductive stability of a molecule or ion. Figure 3.1 indicates that according to MO theory ether solvents are significantly more stable to reduction than ester and carbonate solvents, but the ethers also tend to have a lower oxidative stability than carbonates (Wang et al. 1999). Note, however, that the stability of the solvents to oxidation and reduction is considerably different when they coordinate a cation (e.g., the formation of a coordinate bond involves electron
Fig. 3.1 Molecular orbital energies of aprotic solvents. Reproduced with permission —Copyright 1999, The Electrochemical Society (Wang et al. 1999)
50
3 High Coulombic Efficiency of Lithium Plating/Stripping …
donation, which makes the solvent less/more susceptible to donating/accepting an electron to/from an electrode). For example, it has been demonstrated that the oxidative stability of glymes is greatly increased for highly concentrated electrolytes in which all of the solvent molecules are coordinated (Pappenfus et al. 2004; Yoshida et al. 2011). A corresponding decrease in reductive stability upon coordination is also expected (Aurbach and Gottlieb 1989). Aurbach et al. (1997) used FTIR spectra (Fig. 2.11) to analyze the composition of SEI layers formed on Li electrodes prepared and stored for three days in EC-DMC solutions of 1 M LiAsF6, LiPF6 and LiBF4. They have deduced and listed the possible components on the Li surface and their corresponding peaks in FTIR spectra (Aurbach et al. 1990a, 1994b, 1995c). Some possible chemical reactions on the Li surface have also been proposed as shown in Scheme 3.1) (Aurbach et al. 1990a, 1997).
Scheme 3.1 Major surface reactions of solvents and salts with Li and Li-C, which form surface films. Reproduced with permission—Copyright 1997, Elsevier (Aurbach et al. 1997)
3.2 Electrolyte and In Situ Formed Solid Electrolyte Interphase
3.2.1.1
51
Esters
Esters (Fig. 3.2) are attractive electrolyte solvents due to their high oxidative stability and the very high conductivity of ester-based electrolytes, especially at low temperature. In addition to Li+ cation solvation via the ester carbonyl oxygen, when methyl formate (MF) is used, anion solvation via hydrogen bonding to the MF formyl proton also occurs which further reduces ionic association interactions (Plichta et al. 1987; Venkatasetty 1975; Tobishima et al. 1990). Rauh and Brummer reported that slow gas evolution occurs when Li is contacted with MF or methyl acetate (MA) and a white solid appears on the surface of the Li, whereas n-butyl formate (BF) was inert under the same conditions. LiAsF6 addition improved the Li stability at elevated temperature, and the amount of gas evolution from Li immersed in LiAsF6-ester electrolytes (heated up to 74 °C) indicated that the stability increased in the order MF > MA > ethyl acetate (EA) (Rauh and Brummer 1977b; Dampier and Brummer 1977). In general, the reactivity with Li increased with increasing alkyl chain length (methyl to butyl) and on transitioning from formates to acetates to propionates (Herr 1990). Tobishima et al., however, instead found that a 1.5 M LiAsF6-MA electrolyte had a higher Li cycling CE than a 1.5 M LiAsF6-MF electrolyte—which is also the case for the electrolytes with mixtures of these solvents and EC (Tobishima et al. 1989, 1990, 1995). The reason for this disagreement is unclear. Electrolytes with MA corrode Li to a much greater extent than those with PC, possibly due to the higher solubility of the reaction products resulting in the continuous exposure of Li to the solvent (Rauh and Brummer 1977b). It was found, however, that electrolytes in which MF or MA is mixed with a carbonate solvent such as DMC, DEC, or EC tend to have an improved CE (Tobishima et al. 1990; Plichta et al. 1989; Tachikawa 1993). Herr (1990) indicated that esters may exist in both a stable keto form, as well as a reactive enol form. The enol form may readily be reduced by Li. Since MF does not
Fig. 3.2 Structures of selected aprotic solvents MF
MA
GBL
EA
DMC EC
DEC
PC
Et2O
EMC
THF
DOL
G1
2MeTHF
THP
G2
3 High Coulombic Efficiency of Lithium Plating/Stripping …
52
form an enol, it cannot undergo such a reaction. Thus, MF is relatively stable to Li —in contrast to MA (Table 3.1.) (Herr 1990). Rauh (1975) indicated, however, that degradation may instead occur in a similar manner to the well-known reduction of carbonyl compounds by alkali metals. It is noteworthy that MF is susceptible to hydrolysis from trace amounts of H2O (in both acidic and basic solutions) (Marlier et al. 2005), forming formic acid, which can then be dehydrated to form H2O and CO; the H2O from this then reacts with additional MF. Residual acidic LiAsF5OH in LiAsF6 is believed to have catalyzed this reaction for MA-based electrolytes in early studies. The use of highly purified LiAsF6 significantly decreased this reaction (Herr 1990). Hydrolysis can also occur in LiBF4-MF electrolytes, but it was suggested that the LiF present with the LiBF4 prevents the subsequent dehydration of the formic acid (Herr 1990). Thus, electrolytes with both LiAsF6 and LiBF4 in MF are more stable to Li than those without the latter salt as indicated by improved cycling performance and by an increase in the time (from 1 to 10 months on storage of Li in the electrolyte at 74 °C) before a detectable amount of gas generation was identified (Herr 1990; Plichta et al. 1989; Honeywell 1975; Ebner and Lin 1987). Instead of degrading to H2O and CO, the formic acid may also directly react with the Li since spectroscopic evaluations of the SEI indicate that the major MF reduction product is Li formate (Table 3.1) (Aurbach and Chusid (Youngman) 1993; Ein-Eli and Aurbach 1996). Cyclic esters, such as gamma-butyrolactone (GBL), have a high flash point and low vapor pressure which reduces the flammability of electrolytes with such solvents (Hess et al. 2015). The use of 1 M LiX-GBL electrolytes (with LiAsF6, LiClO4, LiBF4, or LiCF3SO3) resulted in a relatively low Li cycling CE and the SEI formed in these electrolytes did not effectively passivate Li foil from continuous reactions on storage (Sazhin et al. 1994). This may be due to the relatively high solubility of the GBL reaction products (Aurbach and Gottlieb 1989; Aurbach et al. 1991a; Rendek et al. 2003; Aurbach 1989a, b). Kanamura et al. (1995d) noted that when Li foil was immersed in a 1 M LiBF4-GBL electrolyte for 3 days, the electrode resistance initially decreased, but then increased steadily with time and the resulting SEI consisted predominantly of LiF with only a small amount of organic compounds present. The SEI formed on Li foil with GBL electrolytes with LiAsF6, LiPF6, or LiClO4 also contained little to no organic components (Kanamura et al. 1995b). This may again be due to the high solubility of the organic salts produced from GBL degradation (Table 3.1).
3.2.1.2
Alkyl Carbonates
As is the case for esters, alkyl carbonates (Fig. 3.2) are not overly stable to Li reduction (Fig. 3.1). Cyclic carbonates (e.g., EC and PC) are either reduced to Li2CO3 and ethylene (or propylene) or the carbonyl group is instead reduced to form a radical anion which then further reacts to form organic salts and polymers
THF
PC
EC
PMC
EMC
DEC
DMC
Aurbach et al. (1992), (1994a) Aurbach et al. (1987), (1994a) Aurbach et al. (1988)
Ein Ely and Aurbach (1992) Aurbach and Gottlieb (1989), Aurbach (1989a) Aurbach et al. (1994a) Aurbach et al. (1987), (1995b) Ein-Eli et al. (1996) Ein-Eli et al. (1997)
MF
GBL
References
Solvent
ROLi, CH3(CH2)3OLi
LiOH
Li oxides
Li2CO3 (continued)
LiF
ROCO2Li
ROCO2Li
CH3CH(OCO2Li) CH2OCO2Li + propylene
LiF+ H2CO3
Li2CO3
LiF+ ROCO2H
LiF
LiF
HF
CH3OLi, CH3OCO2Li, CH3CH2CH2OLi, CH3CH2CH2OCO2Li, (CH2OCO2Li)2 + ethylene
ROCO2Li
Li2CO3
Li2O Li2O2
RCOOLi + Li2CO3
HCO2Li + Li2CO3
Li oxides ROCO2Li species
Li oxides + HCO2Li
CO2
LiOH-Li2O
(Reaction of ROCO2Li + H2O)
LiO(CH2)3CO2Li
HCO2Li
Contaminants/additives O2 H2O
CH3OLi, CH3OCO2Li
CH3CH2OCO2Li + CH3CH2OLi
ROCO2Li (CH3OCO2Li)
CH3(CH2)2COOLi, cyclic di-Li b-keto ester salt
mostly HCO2Li, ROLi (CH3OLi)
Dry
Table 3.1 Reaction products of solvents, salts, and contaminants with Li (Aurbach 1999)
3.2 Electrolyte and In Situ Formed Solid Electrolyte Interphase 53
Malik et al. (1990) Aurbach et al. (1990a, b) Aurbach et al. (1988), Aurbach and Granot (1997) Aurbach and Granot (1997) Aurbach and Granot (1997) Aurbach et al. (1995a) Ein-Eli and Aurbach (1996) Ein-Eli and Aurbach (1996) Ein-Eli and Aurbach (1996) Ein-Eli and Aurbach (1996)
2MeTHF
MF-PC
MF-EC
MF-DEC
MF-DMC
THF-2MeTHF
G2
Et-DME
DME
DOL
References
Solvent
Table 3.1 (continued)
HCO2Li, ROCO2Li species
CH3OLi, CH3OCH2CH2OLi, (CH2OLi)2 THF reduction products dominate HCO2Li dominates, ROCO2Li (minor)
LiOH-Li2O
Li2CO3, HCO3Li
LiO(CH2)3CO2Li
HCO2Li, Li2O-LiOH
Li alkoxides
CH3CH2OLi (CH2OLi)2 poly DOL species ROLi (CH3OLi)
CH3CH2OLi (CH2OLi)2
Li2O
HCO2Li
ROCO2Li
ROLi species
Contaminants/additives O2 H2O
Li pentoxides
Dry
Li2CO3
HCO2Li
Li2CO3 + red. products
Li2CO3 + red. products
ROLi species
ROCO2Li
CO2
(continued)
LiF
LiF
HF
54 3 High Coulombic Efficiency of Lithium Plating/Stripping …
Ein-Eli and Aurbach (1996) Aurbach et al. (1996) Ein-Eli et al. (1996) Aurbach et al. (1995b) Aurbach et al. (1992) Aurbach and Gofer (1991)
MF-ethers
EC or PCethers
EC-PC
EC-DEC
EC-EMC
EC-DMC
References
Solvent
Table 3.1 (continued)
ROCO2Li species dominate, ROLi (minor)
EC reduction products dominate
HCO2Li dominates
Dry
ROCO2Li, Li2CO3 LiOH-Li2O
Li2O-LiOH
ROCO2Li
Li2CO3, ROCO2Li
Li2O, ROCO2Li
Li2O
Contaminants/additives O2 H2O
Li2CO3, ROCO2Li
ROCO2Li
Li2CO3 + red. products
CO2
LiF
HF
3.2 Electrolyte and In Situ Formed Solid Electrolyte Interphase 55
56
3 High Coulombic Efficiency of Lithium Plating/Stripping …
(Gachot et al. 2008; Herr 1990; Nazri and Muller 1985a; Aurbach et al. 1987; Aurbach and Gottlieb 1989; Ota et al. 2004c; Wang et al. 2001). EC therefore is thus reduced principally to lithium ethylene dicarbonate (CH2OCO2Li)2 and lithium ethoxide (CH3CH2OLi), as well as other degradation products such as (CH2CH2OCO2Li)2, LiO(CH2)2CO2(CH2)2OCO2Li, Li(CH2)2OCO2Li, and Li2CO3 (Wang et al. 2001). PC undergoes similar reactions. PC (and EC) can also be nucleophilically attacked by OH− from the native film (LiOH) or generated by the reaction of Li with trace H2O yielding lithium alkyl carbonate(s) (e.g., ROCO2Li) and Li2CO3 (Table 3.1) (Aurbach 1999). Kanamura et al. noted that when Li foil was immersed in a 1 M LiBF4-PC electrolyte for 3 days, the electrode resistance initially decreased, but then increased steadily with time, and the resulting SEI consisted predominantly of LiF with only a small amount of organic compounds present—as with GBL electrolytes—but the impedance was an order of magnitude higher than those for the comparable GBL electrolytes (Kanamura et al. 1995b). Koike et al. also found that a 1 M LiClO4-PC electrolyte continuously reacted with the deposited Li, indicating that the SEI layer did not effectively passivate the Li (Koike et al. 1997; Nishikawa et al. 2010; Qian et al. 2015b; Ding et al. 2013a). Ding et al. (2013a) scrutinized the CE and morphology of Li deposited on Cu from 1 M LiPF6-carbonate electrolytes (with PC, EC, DMC, and EMC). The deposits from cyclic carbonates (PC and EC) were clusters of thick, tangled needles, whereas those for the acyclic carbonates (DMC and EMC) were more fibrous and only partially covered the Cu substrate (Fig. 3.3). The surface layer of the SEI films from the PC and EC electrolytes were composed principally of LiF and Li2CO3, with more of the former for the PC electrolyte and more of the latter for the EC electrolyte. PC appears to be more reactive with the LiPF6 salt since this electrolyte became discolored on storage, whereas the others did not. More organic compounds/salts that were less effective at passivating the Li surface were formed in the acyclic carbonates. The average Li cycling CE values for the 1 M LiPF6-carbonate electrolytes were the following: PC (77 %), EC (95 %), DMC (24 %), and EMC (7 %) (Ding et al. 2013a). Other studies have also shown that electrolytes with DMC generally result in a superior CE and surface passivation relative to those with DEC (or EMC) (Ding et al. 2013a; Tachikawa 1993). This despite the fact that lithium alkyl carbonate and lithium alkoxide reduction products are formed for all of these acyclic solvents (Table 3.1). Aurbach et al. (1987) noted that Li metal dissolves completely when immersed in pure DEC due to a lack of passivation, resulting in a brownish solution. Tachikawa (1993) also noted a brown discoloration of the solution during Li deposition (on a Ag electrode) using a 1 M LiClO4-DEC electrolyte. Aurbach et al. (1987) reported that the main reaction products were lithium alkyl carbonates and lithium alkoxides (Table 3.1) with no detectable Li2CO3 evident. This suggests that the partially reduced alkyl carbonate undergoes radical reactions rather than a second electron transfer which would yield the fully reduced species (i.e., Li2CO3) (Gachot et al. 2008). This differs from the cyclic carbonates which do tend to form
3.2 Electrolyte and In Situ Formed Solid Electrolyte Interphase
57
Fig. 3.3 SEM images of Li deposition morphologies and the corresponding energy-dispersive X-ray spectroscopy dot maps of Cu (represented by green color) from 1 M LiPF6-carbonate electrolytes with a PC, b EC, c DMC, and d EMC. Reproduced with permission—Copyright 2013, The Electrochemical Society (Ding et al. 2013a)
Li2CO3, perhaps due to the greater stability of the intermediate radical for these cyclic solvents relative to that for DEC. The poor passivation and discoloration originate because the DEC reduction products—CH3CH2OLi and CH3CH2OCO2Li —are soluble in DEC solutions (Schechter et al. 1999; Ding et al. 2013a; Tachikawa 1993).
3 High Coulombic Efficiency of Lithium Plating/Stripping …
58
3.2.1.3
Ethers
Ethers (Fig. 3.2) tend to be much more stable to reduction than esters and carbonates (Fig. 3.1) (Aurbach and Gottlieb 1989). Thus, the use of ethers as electrolyte solvents usually results in the highest CE when plating/stripping Li. Early work focused on diethyl ether Et2O, THF, and 2MeTHF (Koch et al. 1982; Abraham et al. 1982). Kanamura et al. noted that when Li foil was immersed in a 1 M LiBF4-THF electrolyte for 3 days, the electrode resistance variation with time differed markedly from that for electrolytes with GBL and PC. The resistance remained relatively low for approximately 36 h and then grew rapidly to a value comparable to that for the PC electrolyte. The resulting SEI consisted of both LiF and a relatively large amount of organic compounds. It was suggested that the HF present initially reacted with the native film to form LiF and the THF percolated through the native surface layer to the Li and reacted to form organic products (Kanamura et al. 1995d). LiAsF6-2MeTHF electrolytes, however, result in a significantly higher Li cycling CE than do comparable LiAsF6-THF electrolytes, as well as those with Et2O (Goldman et al. 1980; Abraham and Goldman 1983; Abraham et al. 1982, 1986). It has been proposed that ring-opening reactions are significantly slower with 2MeTHF and this accounts for the improved cycling with this solvent (Goldman et al. 1980; Koch 1979). Glymes behave somewhat differently from the cyclic ethers. Aurbach et al. proposed a mechanism for DME reduction in which a radical anion is formed which is stabilized by a Li+ cation which decomposes to CH3OLi and a methoxy ethyl radical. This then undergoes a second electron transfer producing an additional CH3OLi and ethylene (Table 3.1) (Aurbach et al. 1988, 1993). Unfortunately, when used as single solvents for electrolytes with varying Li salts, glymes often result in crystalline solvates with high melting temperatures—except for electrolytes with dilute salt concentrations (Henderson 2006). There are thus few studies of Li plating/stripping in electrolytes with glyme solvents alone. 1,3-dioxolane (DOL) was found to be particularly useful as an electrolyte solvent. Dominey et al. reported that LiClO4-DOL electrolytes resulted in smooth Li deposits with a very high Li cycling CE (Dominey and Goldman 1990; Shen et al. 1991; Dominey et al. 1991). Electrolytes with high purity were found to form SEI layers consisting of both alkoxides (e.g., CH3CH2OCH2OLi) and LiCl (Aurbach et al. 1990b). LiClO4-DOL electrolytes, however, were found to detonate due to an uncontrolled reaction as the electrolyte rapidly polymerized generating excessive heat (Newman et al. 1980). Replacement of the salt with LiAsF6 and the use of additives such as 2mehtylfuran (2MF) and KOH—which scavenge the trace amounts of acidic species that polymerize DOL—resulted in a dramatic improvement in the cyclability of Li/TiS2 cells Dominey and Goldman (1990; Goldman et al. 1989). Tributyl amine (Bu3N) was also found to be a highly effective stabilizer to limit/prevent the chemical polymerization of the DOL for LiAsF6-DOL electrolytes (Gofer et al. 1992). This formulation of electrolyte components resulted in nodular Li deposition with an exceptionally high Li cycling CE (Gofer et al. 1992;
3.2 Electrolyte and In Situ Formed Solid Electrolyte Interphase
59
Zaban et al. 1996; Dan et al. 1997; Aurbach and Moshkovich 1998; Zinigrad et al. 2004; Aurbach 1999; Aurbach et al. 2002b; Mengeritsky et al. 1996a).
3.2.1.4
Mixed Solvents
Interestingly, mixing two or more solvents is found, in some instances, to result in an improved Li cycling CE relative to comparable single solvent electrolytes. For example, this is the case for MF and acyclic carbonate (Table 3.1) (Salomon 1989; Tachikawa 1993; Plichta et al. 1989) and the addition of DMC or DEC to MF-based electrolytes was found to improve the cycling efficiency of Li||LiCoO2 cells (Plichta et al. 1989). Single solvent electrolyte studies with EC have not been extensively studied due to this solvent’s high melting temperature, but it has been widely used as a co-solvent. The combination of EC and other solvents increases the Li cycling CE, often quite significantly relative to the values for the single solvent electrolytes (including ethers), to average values >90 % (Fig. 3.4) (Tobishima et al. 1989, 1990). The Li surface chemistry of EC-DMC and EC-DEC mixtures is dominated by EC reduction, but this is more significant for Li passivation for the mixtures with DMC because the solvent reduction products are less soluble in DMC than in DEC (Schechter et al. 1999). In particular, Wang et al. found that electrolytes with EC, cyclic ethers, and lithium bis(perfluoroethylsulfonyl)imide (LiBETI) resulted in a good cycling performance, high thermal stability, high conductivity, and the homogeneous deposition of nodular Li with a relatively high CE (>90 %) (Wang et al. 1999, 2000; Xianming et al. 2001; Ota et al. 2004a). After 50 plating/stripping cycles, the interphasial layer was much thinner for the electrolyte with the cyclic ether (tetrahydropyranyl, THP) than for DME, carbonates (DMC and PC), and an ester (GBL) (Fig. 3.5), which accounts for the significant variances in the Li cycling performance with these electrolytes (Fig. 3.6) (Ota et al. 2004a). A detailed analysis of the SEI concluded that the outer layer consisted of ROCO2Li (typical for EC), polymers, and LiF, while the inner layers contained Li2O and carbide species with C–Li bonds (Ota et al. 2004c). The polymers were determined to be lithium
Fig. 3.4 Average Li cycling CE on a Pt substrate for 1.5 M LiAsF6-solvent or LiAsF6-EC/solvent (1:1) electrolytes. Reproduced with permission—Copyright 1990, Elsevier (Tobishima et al. 1990)
60
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Fig. 3.5 SEM images of Li deposited on a Ni substrate in a, f EC/THP (1/1), b, g EC/DME (1/1), c, h EC/DMC (1/1), d, i EC/PC (1/1) and e, j EC/GBL (1/1) electrolytes containing 1 M LiBETI. The left images a–e are the surface morphologies after the initial plating at 0.6 mA cm−2 (0.5 C cm−2). The right images f–j are the cross-sectional morphologies after 50 cycles at 0.6/0.6 mA cm−2 (0.5 C cm−2) for the plating/stripping, respectively. Reproduced with permission—Copyright 2004, The Electrochemical Society (Ota et al. 2004a)
ethylene dicarbonate with ethoxy units suggesting that these polymers result from the reduction of both EC and the cyclic ether. Such polymers with both carbonate and flexible ethoxy segments may create a stable, elastic layer that more effectively
3.2 Electrolyte and In Situ Formed Solid Electrolyte Interphase
61
Fig. 3.6 Li cycling CE for the plating/stripping of Li on a Ni substrate in EC/THP (1/1), EC/DME (1/1), EC/DMC (1/1), EC/PC (1/1) and EC/GBL (1/1) electrolytes containing 1 M LiBETI. The current densities were 0.6/0.6 mA cm−2 (0.5 C cm−2) for the plating/stripping, respectively. Reproduced with permission —Copyright 2004, The Electrochemical Society (Ota et al. 2004a)
passivates the surface (than occurs for EC mixtures with glymes, carbonates, and esters), thus explaining the higher CE and reduced interfacial layer thickness.
3.2.2
Influence of Lithium Salts
Kanamura et al. (1995b) found that immersion of Li foil electrodes in 1 M LiX-GBL (with LiX is LiAsF6, LiPF6, LiClO4, or LiBF4) electrolytes resulted in only limited reactions of the native film with the LiAsF6 and LiClO4 electrolytes with major SEI species being LiOH and Li2CO3 and minor species of LiF (or LiCl) in the outer layer and Li2O in the inner layer. For the LiBF4 electrolyte, the SEI outer layer consisted principally of LiF with LiOH and Li2CO3 present as minor species and LiF and Li2O as major and minor components, respectively, of the inner layer indicating the greater reactivity of the BF4− anion. The SEI film formed in the LiPF6 electrolyte, however, was thinner and more compact than that for the other electrolytes, but the resistance of this film was also much higher (with rapid reactions occurring just after immersion) (Kanamura et al. 1994b, 1995b). Koike et al. (1997) also reported that a 1 M LiPF6-GBL electrolyte formed a thin, compact film on a Ni electrode during Li deposition. Kanamura et al. noted that the outer SEI layer resulting from the LiPF6 electrolyte consisted principally of LiF with organic compounds, LiOH and Li2CO3 present as minor species, while the inner layer was composed of LiF and Li2O. After 3 days of storage in the electrolytes, Li was plated onto the Li foil electrodes. This resulted in tangled dendritic needles for the LiAsF6, LiClO4, and LiBF4 electrolytes, but the electrolyte with LiPF6 seems to instead have formed a compact array of needles (Kanamura et al. 1994b, 1995b). This difference in behavior may perhaps be explained by the H2O content in the electrolyte (reported as Va > ECs/Cs+, then Li deposits on the substrate surface (Fig. 3.19a). Due to the unevenness of the substrate surface or other fluctuations in the system, some Li protuberant tips will unavoidably be formed at the substrate surface (Fig. 3.19b).
82
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Then a higher electron-charge density appears at the newly formed Li protuberance that attracts more Li ions from the electrolyte and deposits more on the tip than on other sites, thus forming the Li dendrite as would happen in the conventional electrolyte. However, in the SHES mechanism, more non-Li Cs+ cations will also be adsorbed on the tip by electrostatic attraction force at the same time (Fig. 3.19c). When the adsorbed Li ions deposit on this tip continuously, the number of Li ions around the tip decreases significantly but more and more non-Li cations that do not electrochemically reduce at this potential will accumulate and surround the tip surface. Then a positively charged electrostatic shield is automatically formed by adsorbed non-Li cations on the Li tip (Fig. 3.19d). Due to the charge repelling and steric hindrance effect of the non-Li cation shield, it is more difficult for Li ions to adsorb and deposit on such a tip. Li ions have to deposit at other sites with lower non-Li cation adsorption ratios. As a result, the continuous Li growth on this tip (dendrite root) is stopped. It is reasonable to believe that there should be a lot of tips forming on the Li surface, but no tips would grow exceptionally as shown in Fig. 3.19e. Eventually, a smooth Li surface could be formed as illustrated in Fig. 3.19f; then this self-healing process will repeat itself. In fact, Fig. 3.19 is a simplified illustration of a Li deposition process with SHES additives. In practice, an SEI layer will always form on the surface of Li metal once it is in contact with electrolyte. Therefore, Cs+ ions will accumulate on the outside of the SEI layer and quickly form an electrostatic shield and prevent amplification of Li dendrite growth. The SHES additive approach is different from those of inorganic additives (including Mg2+, Al3+, Zn2+, Ga3+, In3+, and Sn2+) used in previous studies (Matsuda et al. 1991); (Matsuda 1993). These additive metal ions are consumed during each Li deposition and their effectiveness will soon fade with increasing cycles. However, in the proposed SHES mechanism, the additive cations will not be consumed and can last for long-term cycling. This prediction has been verified by chemical analysis of the deposited films. Figure 3.20 shows SEM images of the deposited Li films on copper substrates with different Cs+ concentrations. In the control electrolyte without the CsPF6 additive, Li dendrite formation is clearly observed in the deposited Li film (see Fig. 3.20a). Even at very low Cs+ concentrations (0.005 M), Li dendrite formation is significantly decreased (Fig. 3.20b). When the Cs+ concentration is increased to 0.05 M, the dendrite formation is completely eliminated, resulting in a very distinct improvement (Fig. 3.20c). Further investigation indicated that even if a dendritic Li film was initially formed during Li deposition, it could be smoothened if further deposition was conducted in the electrolyte containing Cs+ additive. In addition to Cs+, Rb+ also exhibits a lower effective reduction potential when its concentration is much lower than that of Li+. Not surprisingly, RbPF6 was also able to suppress Li dendrite growth although it is not as effective as CsPF6. Unlike other cations, reported in the previous work, that will form part of the SEI layer (Matsuda et al. 1991) (Matsuda 1993), SHES additives (Cs+, Rb+, etc.) do not form part of the SEI layer as verified by several trace analysis techniques (Ding et al. 2013b).
3.2 Electrolyte and In Situ Formed Solid Electrolyte Interphase
(a)
(c)
(b)
20 μm
83
20 μm
20 μm
Fig. 3.20 SEM images of the morphologies of Li films deposited in electrolyte of 1 M LiPF6/PC with CsPF6 concentrations of: a 0 M, b 0.005 M, c 0.05 M, at a current density of 0.1 mA cm−2. Reproduced with permission. Copyright 2013, American Chemical Society (Ding et al. 2013b)
Finally, it is necessary to indicate that although the SHES mechanism has addressed the dendritic morphology problem encountered during Li deposition, the CE of Li deposition in the electrolytes used in the above work is still relatively low (*76 % in case of 1 M LiPF6/PC). Further optimization of electrolyte solvent, salt, and additives have increased CE of Li deposition/stripping to more than 98 % and still retained dendrite-free morphology of the deposited Li films, which is required for long-term cycling operation of Li metal anodes.
3.3
Ex Situ Formed Surface Coating
Rather than using the electrolyte components to dictate the composition of in situ formed SEI layers on the Li, a wide variety of ex situ surface treatments and polymer coatings have been applied to the Li anode prior to cell assembly. Examples include: N2 gas (generating Li3N as noted above) (Wu et al. 2011), cyclopentadienyldicarbonyl iron (II) silanes (Fp-silanes) (Neuhold et al. 2014), Al2O3 (Kazyak et al. 2015), carbon thin films (amorphous, nanostructured, and diamond-like) (Zhang et al. 2014c; Bouchet 2014; Zheng et al. 2014; Arie and Lee 2011), Nafion (Song et al. 2015), polyvinylidene fluoride (PVDF)-Li2CO3 (Chung et al. 2003), PVDF-HFP (Jang et al. 2014; Lee et al. 2006; Osaka et al. 1999), PVDF-HFP-Al2O3 (Lee et al. 2015), polyacetylene (Belov et al. 2006), poly (vinylene carbonate-co-acrylonitrile) (P(VC-co-AN)) layer (Choi et al. 2013), plasma-polymerized 1,1-difluoroethene (Takehara et al. 1993), and crosslinked Kynar 2801-1,6-hexanediol diacrylate semi-IPN (interpenetrating polymer network) (Choi et al. 2003, 2004a, b). This last coating transformed the Li from a needle-like deposit to a nodular morphology (Fig. 3.21)—despite the use of the same electrolyte [i.e., 1 M LiClO4-EC/PC (1/1)]—as the amount of the crosslinking agent (i.e., 1,6-hexanediol diacrylate) was increased. Recently, Zheng et al. designed an interconnected hollow carbon nanospheres for stable Li metal anodes (Zheng et al. 2014). They developed a template synthesis
84
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Fig. 3.21 SEM images of Li deposited from a 1 M LiClO4-EC/PC (1/1) electrolyte a without a surface coating, b with a protective coating of only Kynar 2801 and with semi-IPN surface coatings with Kynar 2801/1,6-hexanediol diacrylate ratios: c (7/3), d (5/5) and e (3/7). Reproduced with permission—Copyright 2004, Elsevier (Choi et al. 2004b)
method for fabricating the hollow carbon nanopheres, using vertical deposition of polystyrene nanoparticles (Fig. 3.22a). A colloidal multilayer opal structure is formed on Cu foil by slowly evaporating a 4 % aqueous solution of carboxylated polystyrene particles. The highly monodisperse polystyrene nanoparticles form a close-packed thin film with long range order (Fig. 3.22b). The polystyrene nanoparticles are coated with a thin film of amorphous carbon using flash-evaporation of carbon fibers (Fig. 3.22c). The samples are then heated in a tube furnace to 400 °C under an inert atmosphere, forming hollow carbon nanopheres on the Cu substrate (Fig. 3.22e). TEM characterization shows that the carbon wall has a thickness of 20 nm (Fig. 3.22f). The hemispherical carbon nanospheres are interconnected to form a thin film, which can be peeled off the Cu surface easily. Mechanical flexibility is important in accommodating the volumetric
3.3 Ex Situ Formed Surface Coating
85
Fig. 3.22 Fabrication of hollow carbon nanosphere-coated electrode. a Fabrication process for the hollow carbon nanosphere-modified Cu electrode. Left to right Polystyrene nanoparticles are first deposited onto the Cu substrate; a thin film of amorphous carbon is coated on top of the polystyrene array using flash-evaporation of carbon cord; thermal decomposition of the polystyrene template results in the formation of interconnected hollow carbon nanospheres. b, c, SEM images of the carbon-coated polystyrene nanoparticle array at low (b) and high (c) magnifications. The slight morphology change of the carbon nanospheres to a hexagonal shape could be due to the elevated temperature during the carbon-coating process. d Digital camera image of the as-fabricated hollow carbon nanosphere thin film after removal of the polystyrene template. e Cross-sectional SEM image of the hollow carbon nanospheres. f TEM image of the hollow carbon nanospheres, with wall thickness of 20 nm. g SEM image of the hollow carbon nanosphere thin-film peeled off the Cu substrate. Red dashed line trace of the curvature of bending. Reproduced with permission—Copyright 2014, Nature Publishing Group (Zheng et al. 2014)
change of Li deposition and dissipating the stress exerted on the Li protection layer during cycling. A digital camera image (Fig. 3.22d) and SEM image (Fig. 3.22g) show that the carbon nanosphere thin film can achieve a bending radius of 20 m. They demonstrate that coating the lithium metal anode with such a monolayer of interconnected amorphous hollow carbon nanospheres helps isolate the Li metal depositions and facilitates the formation of a stable SEI. The CE of Li cycling with such an ex situ formed surface layer is 99 % for more than 150 cycles. This result indicates that nanoscale interfacial engineering could be a promising strategy to tackle the intrinsic problems of lithium metal anodes.
86
3.4
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Mechanical Blocking and Solid Electrolytes
Commercial separators for Li-ion batteries tend to be thin ( 16,000 before crosslinking). Unplasticized polyether electrolytes are generally found to exhibit a low Li/SPE interfacial resistance that remains relatively stable over time, even upon storage at elevated temperature (Appetecchi et al. 1998; Appetecchi and Passerini 2002; Fauteux 1993). Importantly, upon discharging the Li anode, the passivation layer formed at this interface is disrupted as indicated by the Li plating/stripping overvoltage dropping to a low value (Appetecchi and Passerini 2002). The ionic resistance of this layer, however, increased substantially with decreasing temperature ( 65 °C) due to the crystallization of some of the PEO (Tm * 65 °C) (Appetecchi and Passerini 2002). This prevents self-discharge and other parasitic reactions during prolonged storage at ambient temperature, while the rapid disruption of the layer after a few initial cycles at elevated temperature enables the cells to perform well even after prolonged storage periods. The cell assembly procedure (including hot lamination) used for the Li/SPE interface was found to be critical for determining the stability of the interface over time (Appetecchi et al. 2000). In situ attenuated total reflectance FTIR measurements have indicated that
3.4 Mechanical Blocking and Solid Electrolytes
87
electrodeposited Li reacts with the PEO in a LiClO4-PEO electrolyte to form alkoxides (ROLi) (Ichino et al. 1991), similar to the reaction of Li with glymes (Aurbach and Granot 1997) and THF: CH2 OCH2 þ Li ! CH2 OLi þ other products After an initial fast reaction period, further reaction with polyethers is relatively slow (Lisowska-Oleksiak 1999). For SPEs containing LiBETI, the interfacial resistance was found to proportionally depend upon the salt concentration (at open circuit) (Appetecchi and Passerini 2002). This indicates that the salt may have a key role in the electrode passivation (either by reacting directly with the Li metal or influencing the reaction of the PEO with the Li). Evaluations of Li surfaces exposed to polyether network polymers (NPs) containing either LiTFSI or LiBF4 indicated that the Li-polymer electrolyte interfacial resistance decreased and then became relatively stable for the electrolyte with LiTFSI, but the resistance grew continuously and became very large for the LiBF4-containing electrolyte (Ismail et al. 2001). XPS indicated that the SEI layers formed in the LiTFSI electrolyte were relatively thin with TFSI− components present at the solution interface (either due to adsorbed or degraded anions), while the layers formed with the LiBF4 electrolyte were thicker and contained much more LiF (Fig. 3.23) (Ismail et al. 2001). These results are analogous to what is obtained for liquid ether electrolytes in which the high stability of the ether solvent results in SEI compositions that are governed by the anion reactions. Contrary to what was initially believed, Li dendrites do form in polymer electrolytes (Fig. 3.24a–c) (Brissot et al. 1998, 1999c, d; Dollé et al. 2002; Liu et al. 2010a, b, 2011b; Wang et al. 2012). To improve the ionic conductivity and restrict
Fig. 3.23 Schematic illustration of the surface film formed on Li foil after contact with a LiTFSI-NP and b LiBF4-NP polymer electrolytes. Reproduced with permission—Copyright 2001, Elsevier (Ismail et al. 2001)
88
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Fig. 3.24 Optical images of Li dendrite growth at a current density of 0.5 mA cm−2 at 60 °C. First column Li|LiTFSI-P(EO)18|Li cell at times of a 0, b 15, and c 20 h second and third column Li|P(EO)18-LiTFSI-1.44PI13TFSI|Li cell at times of d 0, e 30, f 35, g 45, h 65, and i 75 h. Reproduced with permission—Copyright 2010, The Electrochemical Society (Liu et al. 2010b)
Li dendrite growth in polymer electrolytes, the polymers have been crosslinked (Khurana et al. 2014; Ueno et al. 2011), inorganic particle fillers have been added (Appetecchi et al. 1998, 1999; Zhang et al. 2002; Liu et al. 2010a; Kim et al. 2013c) and ILs have been incorporated (Shin et al. 2003, 2005a, b, 2006; Cheng et al. 2007; Choi et al. 2007b, 2011; Kim et al. 2007a, 2010a Zhu et al. 2008; Appetecchi et al. 2010, 2011; Balducci et al. 2011; Yun et al. 2011; Kim et al. 2012; Wang et al. 2012; Wetjen et al. 2013; Swiderska-Mocek and Naparstek 2014; de Vries et al. 2015). Li cycling CE values >95 % were reported for cells with SPEs with and without fillers, and the cells could be cycled for many hundreds of cycles at elevated temperature (e.g., 90 °C) (Appetecchi et al. 1998, 1999, 2000; Appetecchi and Passerini 2002; Gauthier et al. 1985a). The addition of ILs to SPEs enabled the operating temperature to be lowered to 30–40 °C and reduced the tendency for Li dendrites to form and penetrate the SPE (Fig. 3.24d–i) (Shin et al. 2003, 2005b; Liu et al. 2010b). The inclusion of both an inorganic filler and IL did enable repeated Li plating/stripping cycles without dendrite formation (Liu et al. 2011b).
3.4 Mechanical Blocking and Solid Electrolytes
89
To further increase the shear modulus of the SPEs to prevent dendrite propagation, block copolymers have been studied in which a polystyrene component creates a rigid structure, while a PEO component dissolves a Li salt to create an ionic transport percolated pathway. These were first reported by Niitani et al. for polystyrene (PS)-PEO-PS (ABA) block copolymers in which the PEO segments were grafted to a PS backbone forming a percolated network structure (Fig. 3.25) (Niitani et al. 2005a, b, 2009). This was later pursued by Balsara et al. and Epps et al. using PS-PEO (AB) diblock copolymers (Singh et al. 2007; Young et al. 2008, 2011, 2014; Wanakule et al. 2009; Teran and Balsara 2011; Young and Epps 2012; Stone et al. 2012; Teran et al. 2012; Hallinan et al. 2013; Gurevitch et al. 2013; Schauser et al. 2014; Devaux et al. 2015a, b). Such polymers have been dubbed SEO (i.e., styrene-ethoxy) SPEs. Since the conductive phase is only a fraction of the material, SEO electrolytes have a lower conductivity than comparable LiX-PEO electrolytes and, in addition, the PEO segments may crystallize (as PEO or as high melting LiX-PEO crystalline phases), (Young et al. 2008) thus requiring that cells with SEO SPEs be operated at elevated temperature ( 80 °C), and slow Li plating rates are required (Devaux et al. 2015b). It was reported that full Li|SEO|LiFePO4 cells had an efficiency of >99 % for slow discharge rates (for the cathode) (Hallinan et al. 2013; Devaux et al. 2015a, b), but an examination of cycling voltammetries (CVs) for the Li plating/stripping indicates that significant current was consumed for reduction reactions well before 0 V versus Li/Li+ during the cathodic scans, with only a small portion of this recovered during the subsequent anodic scans (Devaux et al. 2015b). In addition, the charge associated with the Li plating (below 0 V vs. Li/Li+) was much higher than that for stripping suggesting that (at least for the CV measurements) the CE for the Li cycling may be quite low (Devaux et al. 2015b). Although such SPEs do restrict dendrite formation, they do not prevent dendrites which ultimately do penetrate the electrolyte, resulting in the short circuit (low impedance) failure of the cells (Fig. 3.26) (Stone et al. 2012; Hallinan et al. 2013; Gurevitch et al. 2013; Schauser et al. 2014). It was noted that this failure mode was common for Li|SEO|Li symmetric cells, whereas capacity fade due to delamination of the Li from the SEO SPE instead occurred for
Fig. 3.25 Schematic illustration of an SEO block copolymer. Reproduced with permission— Copyright 2005, Elsevier (Niitani et al. 2005a)
90 Fig. 3.26 a Typical voltage versus charge passed prior to a short circuit (Cd) data showing the short circuit of the cell at 187 C cm−2 for a Li|SEO|Li cell cycled at 90 ° C with a current density of 0.17 mA cm−2. b Cycling data showing Cd as a function of PEO molecular weight. c Cd as a function of storage modulus, G’ (modulus determined for polymer without Li salt at 10 rad s−1 and 90 °C). Symbols correspond to SEO electrolytes (Filled diamond) and PEO electrolytes (Filled square). Reproduced with permission—Copyright 2012, The Electrochemical Society (Stone et al. 2012)
3 High Coulombic Efficiency of Lithium Plating/Stripping …
3.4 Mechanical Blocking and Solid Electrolytes
91
Li|SEO|LiFePO4 cells (Devaux et al. 2015a). One possible explanation for this is that when the latter cells are cycled, the plating of the Li will result in an expansion of the cell (since the volume increase for the plated Li is significantly larger than the volume decrease for the cathode delithiation). Discharge of the cell then results in a contraction. The repetition of this expansion/contraction during cycling likely results in the delamination, since the SEO SPE is rigid. In contrast, when Li is simply shuttled from one Li electrode to another (i.e., symmetric cell), the cell volume should remain approximately constant during cycling and Li dendrites will extend into the electrolyte from both electrodes through defects in the rigid SPE. Since high modulus (i.e., rigid) block copolymer electrolytes have been unsuccessful at fully preventing dendrites from penetrating across the electrolyte to form short circuits, solid electrolytes with an even higher modulus, such as inorganic (crystalline or glassy) electrolytes, have been developed for batteries with Li anodes.
3.4.2
Solid Inorganic Electrolytes
Inorganic SSEs composed of inorganic compounds are nonflammable and generally more electrochemically stable (Jung et al. 2015; Takada 2013; Fergus 2010; Ren et al. 2015b). Therefore, they have been regarded as the ideal materials to protect Li metal anodes against dendrite penetration because of their unity Li+ transfer number (tLi+) and high mechanical strength, which often far exceeds that of Li itself. Both thin films and bulk forms of solid-state Li-ion conductors have been developed to effectively block Li dendrite growth (Christensen et al. 2012; Shao et al. 2012b). To date, the most widely used inorganic thin-film ion conductor is nitrogen-doped Li-ion phosphate film (LiPON) developed by Bates and Dudney et al. in the early 1990s (Bates et al. 1993). LiPON exhibits a conductivity of 2 10−6 S cm−1 at 25 °C and excellent long-term stability in contact with Li metal. Bates and Dudney also first reported the application of LiPON as a Li-ion-conducting electrolyte and Li metal protection layer in a thin-film battery (Bates 1994). Later on, Herbert et al. (2011) reported that the shear modulus of LiPON is approximately 77 GPa, 7.3 times higher than that of Li, which far exceeds the basic requirement of mechanical strength for electrolyte to suppress the Li dendrites, about twice that of Li. This result is also independent of the substrate type, film thickness, and annealing; therefore, LiPON is expected to be fully capable of mechanically suppressing dendrite formation at the Li/LiPON interface in thin-film batteries. The typical structure of thin-film, solid-state batteries using LiPON as ion conductor is Li|LiPON|LiCoO2 (Bates et al. 1993). The long cycle and shelf lives of these batteries result from the properties of the glassy LiPON electrolyte, which is stable in contact with metallic Li at potentials from zero to nearly 5.5 V and has acceptable conductance in the thin-film form. These batteries have demonstrated a very long shelf life. For example, Dudney (2005) demonstrated that test cells have maintained a charged state for more than a year with negligible loss of capacity
92
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Fig. 3.27 a Capacities recorded for three Li||LiCoO2 thin-film batteries when rapidly discharged after prolonged storage in the fully charged state. b Power and energy density determined from constant-current discharge measurements for thin-film batteries with a Li anode and the indicated thin-film cathode. The batteries were discharged over 4.2–3.0, 4.5–3.0, and 4.5–2.5 V for the cLiCoO2, cLiMn2O4, and nLixMn2−yO4 cathodes, respectively (Filled square). Reproduced with permission—Copyright 2005, Elsevier (Dudney 2005)
(see Fig. 3.27a. This level of reliability is not possible for other types of rechargeable batteries. These batteries also exhibit very high specific energies (Wh/kg) and specific power, as shown in Fig. 3.27b when only the active materials are considered. The effect of substrate and package material on the specific energy and energy density of the batteries will be discussed later. The cycle life of these batteries can exceed more than 40,000 full depth charge–discharge cycles, which far exceeds the cycle lives for other types of batteries. There is no liquid electrolyte, no polymer, or any other organic material present in thin-film, solid-state batteries; therefore, side reactions between electrode and electrolyte are minimized. Many different bulk-form ceramic glass (*50–200 lm thick) Li-ion conductors have also been developed and can effectively suppress Li dendrite growth. One example of these glass electrolytes is LiSICON-type Li1+xAlxTi2−x(PO4)3 (LATP) developed by Fu (1997a, b). Recently, Wang et al. reported a Li||LiMn2O4 battery operated in aqueous electrolyte using a Li metal anode protected by LATP glass. The battery demonstrates excellent stability and good electrochemical performance (Wang et al. 2013c). LATP glass prepared by Ohara glass and other companies has been widely used by many research groups worldwide to protect Li metal glass and has been applied in Li-air and Li-S batteries, as well as other energy storage and conversion systems. LATP glass is stable in weak acid and alkaline electrolyte. One of the disadvantages of LATP glass is that it is not stable when in contact with Li metal. Visco et al. first solved this problem by introducing an interfacial layer (a solid layer such as Cu3N, LiPON, or nonaqueous electrolyte) between the Li metal and the Ohara glass, thus forming a protected Li electrode (PLE) (Visco et al. 2004a, 2009). Figure 3.28 shows the schematic of a PLE proposed by Visco et al. (2009), where the Li electrode was protected by an interfacial layer and a Li metal phosphate glass. Figure 3.26b shows a Li-air battery with a double-sided PLE. Li-air batteries using this PLE can operate in both aqueous and nonaqueous electrolytes (Zhang et al. 2010b; Yoo and Zhou 2011; Li et al. 2012a, b). However,
3.4 Mechanical Blocking and Solid Electrolytes
93
Fig. 3.28 a Use of an interlayer and a water-stable solid electrolyte to protect Li (LMP Li metal phosphate). b Schematic of a Li-air battery based on a PLE technology. Reproduced with permission—Copyright 2009, Elsevier (Visco et al. 2009)
large-scale application of these inorganic solid-state ceramic Li ion conductors is still hindered by their high cost, poor mechanical stability and limited ionic conductivity. Further development of new, flexible, inorganic, solid-state Li-ion conductors with good mechanical strength and stability, high Li ionic conductivity, excellent compatibility with Li metal, and wide electrochemical windows is still under investigation for their application in rechargeable Li metal batteries. Another extensively studied type of solid-state inorganic electrolytes is garnet-like oxide glasses (Li6ALa2Ta2O12 [A = Sr, Ba]) developed by Thangadurai and Weppner (2005; Thangadurai et al. 2003; Murugan et al. 2007). Total conductivity values as high as 1.6 mS cm−1 have been achieved (Du et al. 2015). Also, they have been shown to be stable with Li metal (Cheng et al. 2015a). Although Li dendrites have been observed in garnet oxides, this is likely to be the results of defects (e.g., grain boundaries) in the samples rather than intrinsic mechanical problems of garnet oxides (Ren et al. 2015a; Ishiguro et al. 2013). Kamaya et al. (2011) reported a superionic conductor with a composition of Li10GeP2S12 which exhibited an extremely high Li+ ionic conductivity of 12 mS cm−1. This is much higher than the traditional solid-state electrolytes such as oxide perovskite, La0.5Li1.5TiO3, thio-LISICON, or Li3.25Ge0.25P0.75S4, with ionic conductivities of the order of 10−3 S cm−1. They found highly anisotropic conduction of Li+ ions along one crystal direction and derived low activation energy for ionic conduction in this new material (see Fig. 3.29). Unlike traditional Li3N or LiTi1.7Al0.3(PO4)3, the electrochemical window of Li10GeP2S12 is wide (vs. Li/Li+); thus, it is suitable for many different kinds of anode or cathode materials. When used in an In||LiCoO2 cell, excellent reversibility has been demonstrated with
94
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Fig. 3.29 All-solid-state electrolyte reported by Kamaya et al. Li+ ions transport quickly through partially occupied LiS4 tetrahedra in the Li10GeP2S12 lattice and interstitial positions that are connected by a common edge. Reproduced with permission—Copyright 2011, Macmillan Publishers Limited (Kamaya et al. 2011)
operation voltage at 3.3 V. One disadvantage of Li10GeP2S12, as well as most of sulfur-based Li ionic conductors, is that they are highly sensitive to moisture and unstable when in contact with Li metal. Therefore, they cannot be used in direct contact with Li anodes and their processing requires a very low humidity environment. Since they are even more sensitive to moisture than Li metal itself, so they often need to be processed in an Ar-filled glove box rather than in a dry room, which is suitable to handle Li metal. Although solid, inorganic electrolytes have many advantages, including suppressing dendrite, stability with Li anodes, reducing battery capacity losses from cycling (i.e., a longer cycle life), usability at elevated temperature, improved safety, and high reliability, while other promoted features include the simplicity of cell design, an absence of leakage, a better resistance to shocks/vibrations and aggressive environments, and superior electrochemical, mechanical, and thermal stability; (Fergus 2010; Cao et al. 2014; Wang et al. 2015c; Knauth 2009), seldom does a solid-state electrolyte exhibits all these advantages at the same time. Therefore, the practical applications of the solid-state electrolyte are often limited to thin films and associated with other limiting factors. In thin-film solid-state batteries, the glass electrolyte is used only to separate the anode and cathode. The energy density of solid-state batteries is still very limited because the thickness of electrode is limited by its own ionic and electronic conductivity. Through improvements in the ionic conductivity and stability of SSEs, these electrolytes also may be used as an ionic conductor inside thick electrode layers. Some SSEs have been used with limited success to prepare solid-state batteries (Nishio et al. 2009). In addition to good ionic conductivity of the electrolyte itself, several other conditions need to be optimized for practical applications
3.4 Mechanical Blocking and Solid Electrolytes
95
of high-energy density miniature energy-storage devices. These additional conditions include minimizing the contact resistance between the solid electrolyte and the electrode particles, reducing the amount of SSE used in the batteries, and improving the compatibility of the electrolyte and electrode materials. Most of the existing SSEs cannot satisfy all of these conditions at the same time as analyzed below: 3. Electrolyte/Electrode Stability. Most highly conductive SSEs have limited electrochemical stability windows. That is, they are either unstable against a highly oxidative cathode or against highly reductive anodes, as in the case of La0.56Li0.33TiO3, which involves the reduction of La0.56Li0.33TiO3, the reduction of Ti4+ ions accompanied by the insertion of Li+ ions. On the other hand, a sulfide-based electrolyte (such as 70Li2S 29P2S5 P2S3) and Li1.5Al0.5Ge1.5(PO4)3 is not stable with a Li anode. 4. Large Interface Impedance between Electrolyte and Electrode. Unlike in the case when liquid or polymer electrolytes are used, there is no intimate contact between the electrode and the solid-state electrolyte. For example, LiCoO2 cathode particles have significantly different particle sizes and physical/chemical properties when compared to sulfide-based solid-state electrolytes so ions and electrons are difficult to transfer between the electrode and the electrolyte. Furthermore, this interface impedance may increase with increasing cycle numbers because of mechanical instability (i.e., expansion/shrinkage) of the electrode that occurs during charge/discharge process. 5. Low Capacity and Specific Energy of Full Device. Nagao et al. (2009) and other groups have investigated batteries that are entirely solid state using a sulfide-based electrolyte. However, the cathodes (CuxMo6S8−y) used by Nagao et al. in their batteries have limited capacities (10−3 S cm−1) and a Li+ transference number of unity, the high-rate capability of cells with solid inorganic electrolytes is often inferior to those with liquid electrolytes due to a high contact resistance between the solid electrolyte grains, as well as with the solid electrode materials (Ohta et al. 2006; Sakuda et al. 2010). As a result of these problems, only a few publications are available about the Li cycling CE for solid inorganic electrolytes, as these publications often instead focus on the electrolyte–cathode interface (Ohta et al. 2012, 2013), but what has been published indicates that the CE values are 1 mA cm−2), Li plating leads to Li growth in the pores and along the grain boundaries of the solid electrolyte (Nagao et al. 2013a). This results in crack propagation which facilitates (Bates et al. 1993) Li dendrite growth through the solid electrolyte and the eventual short circuit of the cell (Nagao et al. 2013a). Several reports have confirmed that short-circuiting occurs through grain boundaries and interconnected pores (Sudo et al. 2014; Suzuki et al. 2015; Ren et al. 2015a). Inherent defects within the solid electrolytes from processing may thus limit the ability of these electrolytes to prevent short-circuiting. Even if such initial defects can be eliminated, the mechanical stresses induced during cycling from Li growth and volume changes are likely to generate and grow new defects. Therefore, extensive research still needs to be done before broad commercial application of inorganic solid-state electrolytes.
3.5
Effect of Substrates
The substrate used as a working electrode for the electrochemical reduction of Li+ cations is important (Fig. 3.30) because in many cases the Li+ cations react with the material to form alloys rather than forming Li metal on the substrate’s surface (Dey 1971; Huggins 1988, 1999a, b, 1989a). For non-Li metal working electrodes, the surface layers formed when the metal surface atoms react with contaminants or undergo conversion reactions with Li2O may also complicate the electrodeposition chemistry. Surface roughness is also a key consideration for SEI formation and the way Li deposits on the substrate.
3.5 Effect of Substrates
97
Fig. 3.30 Cyclic voltammograms (CVs) (first cycle) of Ni and Al electrodes with a 1 M LiClO4-PC electrolyte (5 mV s−1). Reproduced with permission —Copyright 1981, Elsevier (Frazer 1981)
3.5.1
Alloys
Alloys are mixtures of metals (or a metal and another element) with metallic bonding. They may consist of solid solutions (single phase) or mixtures of phases. For many Li-based alloys, intermetallic compounds form with specific stoichiometries and unique crystal structures. Li can thus be inserted into other materials and, by avoiding dendritic growth in which the reactive Li metal is directly exposed to the electrolyte, much higher cycling efficiencies can be obtained (Sazhin et al. 1994). Brief descriptions are provided below for the known phase behavior for Li with carbon (graphite) and metals that are commonly used as electrodes.
3.5.1.1
Li-C (Graphite)
Commercial Li-ion batteries are based upon anodes in which Li+ cations intercalate into carbon (typically graphite) to form Li-C alloys (Huggins 2009; Ogumi and Wang 2009; Okamoto 1989). The phase diagram for Li-C(graphite) indicates that multiple phases form: C72Li, C36Li, C18Li, C12Li, C6Li, and aCLi (Okamoto 1989). The progressive intercalation of Li+ cations into graphite during the reduction reaction is referred to as staging, and a key foundation of Li-ion batteries is the reversibility of this reaction up to the C6Li phase (thus stipulating the maximum reversible capacity achievable for a graphite electrode).
3.5.1.2
Li–Cu
Cu, the most commonly used current collector for graphite-based anodes for Li-ion batteries, is frequently used as a working electrode for the plating/stripping of Li metal. Li has a fairly high solubility in Cu when melt is processed at high temperature, mechanically alloyed (i.e., ball milled) or when electrodeposited from molten salts (Baretzky et al. 1995; Gąsior et al. 2009; Pastorello 1930; Camurri
3 High Coulombic Efficiency of Lithium Plating/Stripping …
98
et al. 2003; Lambri et al. 1996, 1999, 2000, 2005; Pérez-Landazábal et al. 2002; Peñaloza et al. 1995). Variable-composition Li-Cu solid solution alloys can thus be formed at elevated temperature up to a composition of approximately 16–18 mol% Li (Pelton 1986c; Gąsior et al. 2009). At lower temperatures, however, there are no indications that Li–Cu solid solutions readily form. Li can thus be plated/stripped on Cu with very high efficiency, although it is possible that some limited reaction occurs between the Li and Cu at the interface, especially during the first Li plating step.
3.5.1.3
Li–Al
Li reacts readily with Al (Armstrong et al. 1989; Besenhard 1978; Besenhard et al. 1985, 1990; Carpio and King 1981; Fischer and Vissers 1983; Frazer 1981; Fung and Lai 1989; Garreau et al. 1983; Geronov et al. 1984a, b, c; Hamon et al. 2001; Huggins 1999b; Jow and Liang 1982; Morales et al. 2010; Myung et al. 2010; Okamoto 2012b; Park et al. 2002; Rao et al. 1977; Suresh et al. 2002; Zhou et al. 2010; Zlatilova et al. 1988; Myung and Yashiro 2014). Figure 3.31 shows that—at 423 °C—Li begins to react above 0.35 V (vs. Li/Li+) to form a solid solution with Al (a phase) up to a concentration of close to 10 atom% (or at.%). This is followed by the formation of the AlLi (b) phase until a concentration just below 50 at.%. This phase is then in equilibrium with the Al2Li3 (s) phase until 60 at.%. Further addition of Li results in a liquid phase (at 150 °C). Note that the potential for the formation of the AlLi intermetallic phase varies linearly with temperature (Huggins 1999b; Wen et al. 1979). Additional intermetallic phases also form at ambient temperature, which further complicates the phase behavior when Li is cycled with an Al electrode. Therefore, Al is not usually used as the current collector for Li deposition.
3.5.1.4
Li–Ni
The phase diagram for the Li–Ni system indicates that the solubility of Li in Ni is extremely small and no intermetallic compounds form (Predel 1997d). Ni electrodes are therefore frequently used for Li plating/stripping studies, although the high cost of Ni tends to preclude its use as a current collector in commercial batteries.
3.5.1.5
Li–Ti
The phase diagram for the Li–Ti system indicates that the solubility of Li in Ti is extremely small and no intermetallic compounds form (Bale 1989d; Predel 1997f). The two metals are almost completely immiscible, even at high temperature with liquid Li.
3.5 Effect of Substrates
99
Fig. 3.31 a Coulometric titration curve for the Li–Al system at 423 °C and b phase diagram of the Li–Al system (the horizontal line corresponds to 423 °C. Reproduced with permission— Copyright 1970, The Electrochemical Society and 2012, Springer (Huggins 1999b; Wen et al. 1979; Okamoto 2012b)
3 High Coulombic Efficiency of Lithium Plating/Stripping …
100
3.5.1.6
Li–W
A phase diagram has not been reported for the Li–W system, but the solubility of Li in W is extremely small and no intermetallic compounds have been identified (Sangster and Pelton 1991a).
3.5.1.7
Li–Pt
Li reacts with Pt to form multiple intermetallic phases (Wibowo et al. 2009; Park et al. 2002; Sangster and Pelton 1991d; Lane et al. 2010b; Bronger et al. 1975, 1985). For an IL-based electrolyte containing LiAsF6, after scanning to negative potential where Li–Pt alloy formation reactions occurred, four different oxidation peaks were noted when the potential was reversed and scanned positively (Wibowo et al. 2009). These peaks varied in magnitude for different temperatures (25–45 °C), indicating that the reactions progressed to a greater extent at elevated temperature and that the relative amounts of the different Li–Pt phases formed were also temperature dependent (Wibowo et al. 2009).
3.5.1.8
Li–Other Metals
No information is available regarding the phase behavior or corrosion characteristics of SS at low potentials with relevant aprotic Li+-based electrolytes, but SS is a common cell component for Li-ion batteries and SS electrodes have been used for Li plating/stripping studies with no indications of reactions occurring between the SS and Li (Kim et al. 2013d). Other metals (such as Al and Pt) that have a complicated phase behavior with Li—with numerous intermetallic phases formed between Li and each metal—include Au (Pelton 1986a; Zeng et al. 2014), Ag (Pelton 1986b), Sn (Sangster and Bale 1998; Bailey et al. 1979), Pd (Predel 1997b; Sangster and Pelton 1992), In (Sangster and Pelton 1991c), Ga (Saint et al. 2005; Yuan et al. 2003; Sangster and Pelton 1991b), Pb (Okamoto 1993; Predel 1997e; Wang et al. 1986), Cd (van der Marel et al. 1982; Pelton 1988, 1991; Wang et al. 1986), and Zn (Okamoto 2012a; Wang et al. 1986) whereas Li has a negligible solubility in Mn (Predel 1997g), Mo (Predel 1997c, 1997h), and Rb (Bale 1989a; Predel 1997a).
3.5.2
Surface Layers and Underpotential Deposition/Stripping
Many studies of Li plating/stripping, especially those involving CV measurements, have identified one or more significant redox events prior to 0 V (vs. Li/Li+) during
3.5 Effect of Substrates
101
cathodic scans which are often attributed to the underpotential depostion/stripping (UPD/UPS) of Li monolayers on Al (Li et al. 1999), Au (Aurbach 1989a; Aurbach et al. 1991b; Chang et al. 2001; Gasparotto et al. 2009; Gofer et al. 1995; Mo et al. 1996; Saito and Uosaki 2003; Wagner and Gerischer 1989; Zhuang et al. 1995; Li et al. 1986), Ag (Aurbach 1989a; Aurbach et al. 1991b), Pt (Aurbach 1989a; Aurbach et al. 1991b; Chang et al. 2001; Paddon and Compton 2007; Lee et al. 2005; Wibowo et al. 2009), Cu (Chang et al. 2001), and Ni (Li et al. 1998a; b; Wibowo et al. 2009, 2010). This phenomenon refers to the reduction of a metal cation on a solid metal surface at a potential more positive than the Nernst potential—the potential at which the metal cation will reversibly deposit on the same metal (i.e., for Li, 0 V vs. Li/Li+) (Sudha and Sangaranarayanan 2002). Note that the assignment of the redox peaks to Li UPD is speculative in many of these publications with limited or no validation provided. Some of these reports mention the possibility of Li alloy formation when using Al, Au, Ag, and Pt working electrodes, and Fig. 3.29a indicates that the equilibrium alloying potentials for the different Li–Al alloys are more positive than 0 V versus Li/Li+. Coulometric titration curves for Li binary systems with Zn, Cd, Pb, Sn, Sb, Bi, In, Ga, and Si reach a similar conclusion (Wang et al. 1986; Huggins 1988, 1999b; Wen and Huggins 1980, 1981), and the alloying potentials are a function of temperature (Huggins 1989b, 1999a). The diffusion of Li within these phases will be a limiting factor for determining which phases form and to what extent. Comparable titration curves are not available for Li binary systems with Al, Au, Ag, and Pt, but it is reasonable to assume that these metals also form their multiple alloy phases at potentials above 0 V versus Li/Li+. Thus, this suggests that the redox peaks often attributed to UPD/UPS reactions are instead due to alloying/dealloying reactions (Tavassol et al. 2013). Why then would redox behavior attributed to UPD be evident on Ni electrodes as well, since Ni does not alloy with Li? Fujieda et al. attributed this to the reactions of the Ni with trace water—which is reduced near 1 V (vs. Li/Li+) (Fig. 3.32.) (Aurbach 1989a; Aurbach et al. 1991b)—to form an electroactive nickel hydroxide electrode surface layer (Fujieda et al. 1998). Publications by Kim et al. may also be relevant to the redox reactions noted for Cu and Ni electrodes (Kim et al. 2013d, 2011). These authors heated Cu and Ni electrodes in air. At 200 °C and above, the Cu surface formed CuO and Cu3O4, whereas at 300 °C and above for Ni, a NiO surface layer formed. These (nonnative) layers significantly increased the cathodic reactions occurring prior to 0 V vs. Li/Li+, which is unsurprising since the oxides NiO (Wang et al. 2010; Needham et al. 2006; Liu et al. 2011a; Li et al. 2010; Kim et al. 2011; Huang et al. 2006, 2009) and Cu2O/CuO (Zhang et al. 2004a; Xu et al. 2015; Wang et al. 2015b; Shu et al. 2011; Grugeon et al. 2001; Gao et al. 2004; Débart et al. 2001) have been used as anodes that undergo conversion reactions as follows: Cu2 O þ 2Li þ þ 2e $ 2Cu0 þ Li2 O þ 2e NiO þ 2Li þ þ 2e $ Ni0 þ Li2 O þ 2e
102
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Fig. 3.32 Schematic view of typical electrochemical processes occurring in Li salt electrolytes with PC, THF, and DME on Au and Ag electrodes. Reproduced with permission—Copyright 1991, Elsevier (Aurbach et al. 1991b)
Given this, it is interesting to note that Aurbach et al. found that the reduction of dissolved O2 occurs near 2 V (vs. Li/Li+) in aprotic electrolytes (Fig. 3.32) (Aurbach 1989a; Aurbach et al. 1991b). In addition to reacting with Li+ at the electrode surface to form Li2O, this reduction of O2 may also form metal oxide surface layers with the respective electrode metals, which then undergo subsequent conversion reactions (and the metals presumably also react with Li2O at the metal-SEI interface). For Al and noble metal electrodes, other oxides such as Al2O3, PtO2, etc., may perhaps also complicate the surface redox reactions. In addition to such considerations, aprotic solvents tend to degrade to form surface layers prior to 0 V vs. Li/Li+ which further obfuscates the redox reactions occurring on varying electrodes prior to Li deposition. This confluence of reactions may explain, in part, the “incubation or initiation period” (which tends to increase with decreasing current density) that is sometimes noted to occur prior to Li plating during galvanostatic charging on metal electrodes, as well as the often observed low CE for the first cycle of Li plating/stripping (Fig. 3.33) (Brissot et al. 1999c; Ota et al. 2003, 2004a; Nishikawa et al. 2007, 2010, 2011; Nishida et al. 2013; Sano et al. 2014).
3.5 Effect of Substrates
103
Fig. 3.33 Li plating on Ni using a 1 M IL-LiTFSI electrolyte after the start of electrolysis depending on a time and b coulomb quantity (open circles indicate when precipitates were first visible in the optical videos). Reproduced with permission —Copyright 2013, Elsevier (Nishida et al. 2013)
3.5.3
Surface Roughness
Dampier and Brummer noted that, when plating Li on a Ni substrate, dendritic Li accumulated as isolated material at the sharp, outer edges of the substrate (Dampier and Brummer 1977). Only a few nodules of Li were evident inside the outer border. When a second cell was constructed with the sharp edges rounded down, however, the dendrites were almost completely eliminated at the edges. Nishikawa et al. also noted that needles tended to form at the edges of a Ni electrode, whereas smaller faceted deposits instead formed in the center of the electrode, indicating that nucleation and needle growth were more rapid at the edges (Nishikawa et al. 2011). This implies that surface roughness is a significant factor for the Li deposition morphology and cycling CE. Morigaki and Ohta (1998) noted that the Li metal surface may contain many grain boundaries, ridge lines, and flat areas. When the Li was stored in dry air, the ridge lines and grain boundaries were covered with Li2CO3 and Li2O. When the Li was then stored in a 1 M LiClO4-PC electrolyte for 24 h, Cl was found at the raised
104
3 High Coulombic Efficiency of Lithium Plating/Stripping …
features, indicating that LiCl formed, whereas the solvent (PC) was degraded (perhaps forming ROCO2Li) on both the raised and flat portions of the surface. After depositing Li (2 mA cm−2 and 0.36 C cm−2) on this Li surface, particles of Li grew preferentially on the raised features and it appeared that the growth of the particles was from the base of the Li. The authors attributed this to enhanced diffusion at the raised features due to the different SEI compositions. The mechanical modification of a Li surface using a micro-needle technique created small indentations in the surface. This increased the surface area of the Li and improved the capacity retention of batteries with these electrodes (Ryou et al. 2015). The Li was found to preferentially deposit in the striated walls of the indentations. The authors attributed this to the removal of the native surface layer on the Li substrate and to reduced current density due to the greater surface area. The results noted above from Morigaki and Ohta, however, suggest that the striated edges may be even more reactive to electrolyte impurities than the flat Li substrate surfaces. Relevant to this work is a study by Gireaud et al. that examined Li deposition on Li substrates that were pitted, cracked, or smooth (Gireaud et al. 2006). When Li was deposited on a pitted (but otherwise smooth) Li surface, dendritic Li deposits formed in the pit-like holes (comparable results were obtained for current densities of 1 and 50 mA cm−2) with no Li deposited on the smooth surfaces (Figs. 3.34a) (Gireaud et al. 2006). Yoshimatsu et al. (1988) noted similar deposition behavior in pits on a Li substrate. When Li was deposited instead on the cracked Li substrate, the Li initially deposited on the crack ridge lines and these deposits then aggregated together to coat portions of the surface (Figs. 3.34b) (Gireaud et al. 2006). The separator constrained the deposits to the interiors of the pit holes. In contrast, when Li was instead plated on polished Li substrates under the same conditions, different deposits formed that had either a small grain (1 mA cm−2) or free-dendritic mossy (50 mA cm−2) (Fig. 3.34c) morphology all over the surface.
Fig. 3.34 Surface morphology after Li plating on different Li surfaces in 1 M LiPF6-EC/DMC: a pitted Li surface; b cracked Li surface (10 s), c polished surface. Inset micrograph magnification showing the free-dendritic mossy deposit spread all over the surface. Reproduced with permission —Copyright 2006, Elsevier (Gireaud et al. 2006)
3.5 Effect of Substrates
105
These factors are well reflected in the cycling characteristics of compressed Li powder anodes (Kim et al. 1988, 2006, 2007b, 2010b; Kwon et al. 2001; Park and Yoon 2003; Hong et al. 2004; Kim and Yoon 2004a, b; Chung et al. 2006; Seong et al. 2008; Kong et al. 2012; Lee et al. 2013). The rougher surface results in a higher surface area (as compared to Li foil), which reduces the effective current density on the anode, which in turn results in improved cycling characteristics (see below) (Park and Yoon 2003). Despite this, the growth in the interfacial impedance upon storage (aging—see below) was lower for the compressed Li powder than for the Li foil (Park and Yoon 2003; Kim and Yoon 2004a; Kong et al. 2012). This is likely due to the presence of a thicker native SEI film on the Li powder anode (than for the Li foil anode)—as reflected by an initially greater impedance for the powder electrode—which may suppress further reactions with the electrolyte to a greater extent (Hong et al. 2004; Kim and Yoon 2004a), as well as differences in reactivity with the electrolyte components due to the surface roughness. Rather than isolated pitting on the flat Li surface during stripping and subsequent deposition inside the pits as observed for the Li foil, the stripping of Li from the powder occurred relatively uniformly across the entire electrode, forming pits within the powder spheres. Deposition then occurred within the pits to reconstruct the compressed spheres (Kim et al. 2006, 2007b; Seong et al. 2008; Kong et al. 2012). During cycling, this resulted in lower electrolyte consumption, less dendritic growth, and reduced interphasial layer formation (Kim and Yoon 2004b; Chung et al. 2006; Kim et al. 2010b).
3.6 3.6.1
Influence of Charge/Discharge Profiles Influence of Pulsed Plating
A coarse grain simulation model for the plating of Li metal—which accounts for the heterogeneous and nonequilibrium nature of the plating dynamics, as well as the long time- and length-scales for dendrite growth—found that dendrite formation increased with an increase in the applied overpotential, and the application of pulsed charging significantly suppressed dendrite formation at high overpotentials (Mayers et al. 2012). According to this model, dendrite growth reflects a competition between the time scales for Li+ cation diffusion/reduction to/at the anode-SEI interface. The use of low overpotentials or short charging pulses favors cation diffusion, thus decreasing the proclivity for dendrite formation. This has been experimentally validated using Li/Li cells with a 1 M LiPF6-EC/DMC (1/1 v/v) electrolyte by comparing the morphology of Li deposited by direct current (DC) and either pulse plating (PP) or reverse pulsed plating (RPP) (Fig. 3.35) (Yang et al. 2014). The PP method with short, widely spaced current pulses improved both the Li morphology (larger Li particles/nodules) and the cycling CE.
106
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Fig. 3.35 Waveforms for DC, PP, and RPP charging, where ia, ic, ta, tc, and toff are the anodic current density (mA cm−2), cathodic current density (mA cm−2), anodic-on time (ms), cathodic-on time (ms) and current-off time (ms), respectively. Reproduced with permission—Copyright 2014, Elsevier (Yang et al. 2014)
In contrast, the RPP method with high current density charge pulses improved the CE, but no improvement in the plating morphology was evident under the test conditions.
3.6.2
Influence of Plated Charge
Sazhin et al. examined the dependence of the CE on the plated charge amount using Li state diagrams (Fig. 3.36) (Sazhin et al. 1997). Different behavior was noted for different electrolytes. For the 1 M LiPF6-EC/DEC electrolyte, the CE was relatively independent of the amount of plated charge up to a value of 0.23 C cm−2, but then the CE deceased with increasing plated charge. For the 1 M LiClO4-EC electrolyte, however, the CE values were much lower and declined with decreasing plating charge. This suggests that the LiPF6 electrolyte passivates the Li much more effectively, but as more Li is deposited, more “dead” Li limits what may be recovered (perhaps due to longer needle-like deposits). In contrast, the LiClO4 electrolyte does not effectively passivate the Li, but the buildup of the degradation products for higher plated charge may shield the deposited Li to some extent from the electrolyte, thus reducing the overall reactivity. Ota et al. plated/stripped Li on a Ni electrode in a 1 M LiBETI-EC/THP (1/1) electrolyte using different amounts of plated charge (Ota et al. 2004a, c). The initial deposits had a fibrous shape, but with increasing applied charge, these grew into larger and larger particle/nodule-like structures (Fig. 3.37). After 50 cycles, the residual interphasial layer thicknesses (after stripping) for the plating/stripping with 0.5, 1.0, and 2.0 C cm−2 plated charge were 11, 26, and 49 µm, respectively (Fig. 3.37), indicating that the thickness of this layer roughly scales with the amount of plated charge. The cycling CE in the initial cycles was highest for the largest plated charge, but comparable CE values were obtained irrespective of the amount of plated charge after 10 or so cycles, and these values slowly decreased upon continuous cycling (Fig. 3.38). The larger particles may be the reason for the higher CE, but it is unclear whether the same particle deposition characteristics are retained after repeated cycling with the buildup of the interphasial layer on the electrode surface.
3.6 Influence of Charge/Discharge Profiles
107
Fig. 3.36 Dependence of CE and capacity loss rate on amount of plated Li for Li plating/stripping on a SS substrate in a 1 M LiPF6-EC/DEC and b 1 M LiClO4-EC electrolytes. Reproduced with permission—Copyright 1997, Elsevier (Sazhin et al. 1997)
3.6.3
Influence of Plating (Charge) Current Density
The plating current density strongly affects the morphology of the deposited Li, as well as its reactivity. The use of different electrolyte compositions (i.e., different salts/solvents/additives, ILs, polymer electrolytes) often results in substantial differences in these Li plating characteristics. The wide variability in experimental
108
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Fig. 3.37 SEM images of Li deposition for variable plated charge for Li plating on a Ni substrate in a 1 M LiBETI-EC/THP (1:1) electrolyte: a–c surface morphology after the initial deposition at 0.2 mA cm−2 and d–f cross-sectional morphology after 50 cycles at plating/stripping current densities of 0.2/0.6 mA cm−2, respectively. Reproduced with permission—Copyright 2004, The Electrochemical Society (Ota et al. 2004a)
Fig. 3.38 Li cycling CE as a function of plated charge for Li plating on a Ni substrate in a 1 M LiBETI-EC/THP (1/1) electrolyte. The plating/stripping current densities were 0.2/0.6 mA cm−2, respectively. Reproduced with permission—Copyright 2004, The Electrochemical Society (Ota et al. 2004a)
3.6 Influence of Charge/Discharge Profiles
109
materials/procedures used in reported studies, however, makes it a challenge to arrive at definitive conclusions about the impact of the plating current density. Sano et al. noted that a higher plating current density results in a higher overpotential and a corresponding larger number of nucleation sites for Li deposition (Fig. 3.39) (Nishikawa et al. 2011). Thus, they observed that the few nuclei present at low plating current density tended to grow into larger particles (for the same total charge passed), while at high current density, needle-like morphologies predominated. Ota et al. plated/stripped Li on a Ni electrode in a 1 M LiBETI-EC/THP (1/1) electrolyte at different plating current densities (with a fixed total charge for each of 0.5 C cm−2) (Ota et al. 2004a, c). The cycling CE increased substantially with decreasing plating current density (Fig. 3.40). SEM images (Fig. 3.41) show that the Li had a particle-like morphology for all of the current densities, but the particles became finer with increasing rate. After 50 cycles, the residual interphasial layer thicknesses (after stripping) for the deposition at 0.2, 0.6, and 1.0 mA cm−2 were 11, 14, and 38 µm, respectively (Figs. 3.37 and 3.41). Thus, despite the same amount of charge for the deposition, the quantity of dead Li increased sizably with increasing plating current density—in agreement with the differing cycling CE values. This is also consistent with the observation reported by Lv et al. (2015).
Fig. 3.39 Schematic of the effect of current density on the distribution of Li nuclei and the morphology of Li deposits. Reproduced with permission—Copyright 2014, The Electrochemical Society (Sano et al. 2014)
110
3 High Coulombic Efficiency of Lithium Plating/Stripping …
Fig. 3.40 Li cycling CE as a function of Li plating current density (0.2, 0.6 and 1.0 mA cm−2) for Li plating on a Ni substrate in a 1 M LiBETI-EC/THP (1/1) electrolyte at 0.5 C cm−2. The stripping current was 0.6 mA cm−2. Reproduced with permission—Copyright 2004, The Electrochemical Society (Ota et al. 2004a)
Fig. 3.41 SEM images as a function of current density for Li plating on a Ni substrate in a 1 M LiBETI-EC/THP (1/1) electrolyte at 0.5 C cm−2: a, b surface morphology after the initial plating at 0.6 and 1.0 mA cm−2 and c, d cross-sectional morphology after 50 cycles at 0.6 and 1.0 mA cm−2 for plating and 0.5 mA cm−2 for stripping, respectively. Reproduced with permission—Copyright 2004, The Electrochemical Society (Ota et al. 2004a)
3.6 Influence of Charge/Discharge Profiles
111
Arakawa et al. (1993) found for Li||amorphous-V2O5 coin cells with a LiAsF6EC/2MeTHF electrolyte that the cycle life increased with a decrease in charge current density. The cells were discharged at 3.0 mA cm−2 and then charged at either 0.5 or 1.5 mA cm−2 resulting in particle-like or needle-like deposits, respectively. In agreement with this, Aurbach et al. (2000) noted that Li||Li0.3MnO2 AA batteries with a 1 M LiAsF6-DOL (stabilized with Bu3N) electrolyte could be discharged at high rates for more than 300 full depth-of-discharge (DOD) charge– discharge cycles, but this required slow charging rates (99 % and only 11 % capacity fade.
Fig. 4.8 Schematic of the hybrid anode design to manipulate the surface reactions on Li–S batteries. Reproduced with permission —Copyright 2014, Macmillan Publishers Ltd. (Huang et al. 2014a)
4.2 Lithium Metal Anodes in Lithium–Sulfur Batteries
163
Besides insertion of a physical barrier between the S cathode and the Li anode, a conformal coating of inorganic or organic materials on the Li metal surface is also an effective strategy to prevent corrosion or side reactions of the Li metal from forming a passivation layer. One example of such coating layers was formed by a crosslinking reaction of a curable monomer in the presence of liquid electrolyte and a photo-initiator (Lee et al. 2003). With the protection layer, an enhanced cell performance was achieved in comparison with cells using a polymer electrolyte. In another work, conductive polymers, such as polypyrrole, have been used to form a protective coating layer on Li powder for Li–S batteries; it can inhibit the side reactions between the Li powder anode and dissolved polysulfides in the electrolyte, minimizing the overcharge and reducing the capacity fading (Oh and Yoon 2014). Similarly, it has been reported that Li powders with protective coatings, compared to Li metal foil, exhibit more advantages, including dendrite suppression, leading to a safer manufacturing process and safer cell operation (Heine et al. 2014). Kim et al. (2013b) reported that a coating layer of a Li–Al alloy could mitigate the polysulfide shuttle phenomenon in Li–S batteries. A protective film consisting of a Li–Al alloy and lithium pyrrolide film was formed on the Li electrode in an electrolyte containing pyrrole and AlCl3 (Wu et al. 2013). The formed protective film was dense and tight with high stability in the electrolyte, giving rise to a low interface resistance and a greatly improved cycling CE.
4.2.2
High Coulombic Efficiency and Dendrite Prevention
In Chap. 3, we have described in detail the approaches to enhance CE and prevent Li dendrites in general Li metal-based batteries. In this section, our review will focus on enhancing CE and preventing Li dendrite growth specifically related to Li– S batteries.
4.2.2.1
Effects of Liquid Electrolytes
Organic liquid electrolytes are some of the foremost electrolytes studied in literature on Li–S batteries (Manthiram et al. 2014; Xu et al. 2014a; Chen and Shaw 2014; Scheers et al. 2014; Yin et al. 2013; Bresser et al. 2013; Song et al. 2013b; Zhang 2013b; Barghamadi et al. 2013), probably due to their easy availability, vast variation, excellent wettability on S cathodes and Li anodes, low viscosity to fill in the micropores of the S cathode, relatively high ionic conductivity, low interfacial resistance, reasonable polysulfide solubility, good chemical stability with Li metal anodes, and a good electrochemical window. However, carbonate solvents (both linear and cyclic) are reactive to reduced polysulfide species (Yim et al. 2013; Gao et al. 2011), and sulfone solvents are also incompatible with polysulfide species, according to the poor discharge and charge voltage profiles (Gao et al. 2011); thus, ether-based solvents, including glymes, are the most commonly used organic
164
4 Application of Lithium Metal Anodes
solvents in Li–S battery studies because of their stability with the Li metal and their compatibility with elemental S cathodes. The Li salts commonly used in Li–S batteries include LiTFSI and LiSO3CF3 (or LiTf) (Zhang 2013b; Scheers et al. 2014). Song et al. (2013a) studied the effect of FEC as a co-solvent in electrolytes on the electrochemical performance of the Li metal anode in Li–S batteries. The Li||Li symmetric cells with the ether-based electrolytes, 1 M LiPF6 in tetra(ethylene glycol) dimethyl ether (Tetraglyme, G4) with and without FEC, showed quite different performances. The Li||Li cell with the control electrolyte without FEC exploded after 4 h charging, while the symmetric cell with FEC-containing G4-based electrolyte lasted for 70 h without shorting. When the ether-based electrolyte was changed to a carbonate-based electrolyte, i.e., 1 M LiPF6 in EC/EMC (3:7 v/v), the Li||Li cells with the control electrolyte without FEC or with 5 % FEC showed unstable potential behavior within 5 to 6 Li stripping/deposition cycles and failed at the 7–8 cycles. With an increase in FEC content in the carbonate solvent mixture (EC/EMC/FEC), 60 % FEC could lead to at least 80 stable cycles, or 1600 h of repeated stable Li stripping/deposition processes, while the control electrolyte showed failure at the 10th cycle. At the same time, the 60 % FEC cell showed extremely low polarization voltage (*5 mV) after 5 Li stripping/deposition cycles, and the cell impedance for the 60 % FEC cell was only about one-third of that for the control cell. This FEC-derived, SEI-protected Li anode gave stable cycling performance in a Li–S battery with a LiTFSI-TEGDME electrolyte, but it showed decaying cycling stability in the Li–S battery with a LiPF6-TEGDME electrolyte because LiPF6 is not stable with the produced polysulfide species (Zhang 2013b). When the Li electrode was pretreated in a 60 % FEC electrolyte, the preformed LiF-based SEI film could slow the migration of soluble polysulfide species to the Li surface so the surface protective layer was smooth, with no physical fractures and very little S detected. On the contrary, the Li electrode pretreated in a G4-based electrolyte showed significant overcharging during the pre-cycle process in a Li–S battery, indicating the shuttle phenomenon of the soluble polysulfides, which is due to the good compatibility of G4 and long-chain polysulfides (Barchasz et al. 2012).
4.2.2.2
Effects of Lithium Salts
Typically, the salt concentration of an electrolyte is less than 1.2 M due to the considerations of ionic conductivity, viscosity, solubility, and cost. Suo et al. (2013) reported a new class of “solvent-in-salt” electrolytes of LiTFSI/DOL-DME with high salt concentrations up to 7 M. Although the ionic conductivity decreased greatly (from about 15 mS cm−1 for 1 M to 0.8 mS cm−1 for 7 M) and the viscosity increased significantly (from about 1 cP for 1 M to 72 cP for 7 M) with the increase in salt concentration, the Li+ transference number increased sharply from about 0.46 for 1–4 M salt concentrations to about 0.73 for 5–7 M. In addition, when this ultrahigh salt concentration (7 M) electrolyte was used in Li–S batteries,
4.2 Lithium Metal Anodes in Lithium–Sulfur Batteries
165
Fig. 4.9 SEM images of Li anodes and Li deposition/stripping profiles: a fresh Li metal, b Li anode after 278 cycles in 2 M LiTFSI/DOL-DME, c Li anode after 183 cycles in 4 M LiTFSI electrolyte, d Li anode after 280 cycles in 7 M LiTFSI electrolyte. The white scale bar represents 60 lm. e, f, g are for electrolytes with 2 M, 4 M, and 7 M salt concentration, respectively. Reproduced with permission—Copyright 2013, MacMillan Publishers Ltd. (Suo et al. 2013)
it showed very little dissolution of polysulfide species, stable capacity (from an initial 1041 mAh g−1 to 770 mAh g−1 at the 100th cycle), high capacity retention (74 % for 100 cycles), high CE (from 93.7 % for the first cycle to nearly 100 % after the second cycle), good rate capacity (1229, 988, 864, 744, and 551 mAh g−1S at current rates of 0.2 C, 0.5 C, 1 C, 2 C, and 3 C, respectively), and good low-temperature performance (about 600, 550, and 350 mAh g−1 at −10, −15, and −20 °C, respectively, at 0.1 C rate). However, the lower salt concentration electrolytes resulted in rapid capacity fade, poor capacity retention, and a large variation of CE. More importantly, the concentrated electrolyte provided a good protection of the Li metal anode, showing the lowest roughness and damage levels compared with the Li anodes in low salt concentrations (Fig. 4.9). This demonstrated that high salt concentration electrolytes could significantly diminish the side reactions between the Li metal and the polysulfide species, effectively reduce corrosion of the Li anode, and suppress Li dendrite formation. The reasons include the inhibition of polysulfide dissolution and fewer free solvent molecules in the concentrated electrolytes. It should be noted that low conductivity and high viscosity of concentrated electrolytes may cause significant increase of impedance of Li–S batteries. Recently, a concentrated electrolyte of 5 M LiFSI was also reported to form
166
4 Application of Lithium Metal Anodes
protective layers on both cathode and anode sides (Kim et al. 2014). Compared to the LiTFSI-based electrolyte, the LiFSI-based electrolyte offered lower solubility of polysulfides, resulting in better performance in Li–S batteries. Miao et al. (2014) used an electrolyte with dual or binary Li salts, LiFSI, and LiTFSI, dissolved in a mixed ether solvent to address the Li dendrite formation and the low Li CE. When Li was deposited on and stripped from a SS substrate, the efficiency of the stripped charge over the deposited charge reached 91.6 % at the first cycle for the dual-salt electrolyte of 0.5 M LiFSI-0.5 M LiTFSI in DOL-DME (2:1 v/v) and became stable at around 99 % after several cycles, which is much higher than those in the conventional LiPF6/EC-DMC electrolyte for Li-ion batteries and the LiTFSI/DOL-DME electrolyte for Li–S batteries (Fig. 4.10a). The LiFSI-LiTFSI dual-salt electrolyte also resulted in excellent long-term Li deposition/stripping performance for the tested 3075 h, and the overvoltages during the deposition/stripping processes were low (7–15 mV) and stable after more than 300 h cycling (Fig. 4.10d). On the contrary, the two control electrolytes showed a much shorter lifetime of repeated deposition/stripping cycles and the LiPF6/
Fig. 4.10 Li deposition/stripping performances in Li|electrolyte|SS cells: a CE versus cycle number for three different electrolytes; b, c, d galvanostatic voltage–time curves in LiPF6/ EC-DMC, LiTFSI/DOL-DME and LiFSI-LiTFSI/DOL-DME, respectively. Each deposition was conducted at 0.25 mA cm−2 for 2.5 h, and the stripping was performed at the same current density with a cutoff voltage of 1.2 V. Reproduced with permission—Copyright 2014, Elsevier (Miao et al. 2014)
4.2 Lithium Metal Anodes in Lithium–Sulfur Batteries
167
EC-DMC electrolyte even had a higher overvoltage, up to about 44 mV (Fig. 4.10b, c, respectively).
4.2.2.3
Effect of Additives
Following successful reports by Mikhaylik (2008) and by Aurbach et al. (2009), LiNO3 has been widely used as a salt additive in electrolytes for Li–S batteries to effectively protect the Li metal anode (Liang et al. 2011b; Zhang 2012a, b, 2013b; Xiong et al. 2011; Scheers et al. 2014; Manthiram et al. 2014). Liang et al. (2011b) reported that the addition of 0.4 M LiNO3 in a 0.5 M LiCF3SO3/DOL-4G (1:1 v/v) electrolyte could lead to quite a smooth and dense Li surface after 20 cycles. A passive film was found coated on the Li anode surface. Xiong et al. (2012) investigated the components and morphologies of the surface films on the Li anode by using LiNO3 as the sole Li salt in an ether-based electrolyte for Li–S batteries. The strong oxidation of LiNO3 led to its reactions with electrolyte components and the Li metal anode, resulting in a homogeneous surface film containing both inorganic species (such as LiNxOy from the direct reduction of LiNO3 by Li metal and Li2SOy from the oxidation of sulfur) and organic species (such as ROLi and ROCO2Li), as Aurbach et al. (2009) indicated. When Li metal was immersed in the LiNO3-containing electrolyte, the Li metal surface became more smooth and compact with a longer immersion time (as shown in Fig. 4.11). This homogeneous and compact surface film would enhance the stability of the Li metal anode and improve the cycle life of Li–S batteries. Besides LiNO3, several other salts and ILs have been reported as electrolyte additives in Li–S batteries to suppress Li dendrite formation. Xiong et al. (2011) studied the effect of LiBOB as an electrolyte additive for electrochemical properties and Li surface morphologies in Li–S batteries. With the increase of LiBOB content in 1 M LiTFSI/DOL-DME (1:1 v/v) from 0 to 10 wt%, the a.c. impedance of the Li anode increased and the discharge capacity of the Li–S battery showed a maximum value at 4 % LiBOB content. However, the cycling stability of Li–S batteries with various LiBOB contents was poor, and nearly no difference could be observed for different LiBOB contents. The authors speculated that the main reasons were attributed to the formation of irreversible Li2S, structural invalidation of the cathode matrix, and the parasitic reactions between polysulfides and the Li metal. The impact of LiBOB on the cycling performance was small. However, the addition of LiBOB in the electrolyte did form smoother and denser surface morphology on the Li anode, and the roughness of this surface morphology reduced as LiBOB content increased. In contrast, without LiBOB, the Li anode surface showed loosely compacted needle-like dendrites after 50 cycles, causing serious safety concerns. Therefore, the addition of LiBOB as an additive or co-salt could prevent the reactions between the electrolyte and the Li metal. Wu et al. (2014) investigated the protection of Li anode and improvement of cycling performance for Li–S batteries by using LiDFOB as an additive in the electrolyte of 1 M LiTFSI/DOL-DME. In the control electrolyte without LiDFOB,
168
4 Application of Lithium Metal Anodes
Fig. 4.11 SEM images of the Li metal surface a as-received, and immersed in 0.5 M LiNO3/ DOL-DME (1:1 v/v) for b 1 h, c 3 h and d 5 h. Reproduced with permission—Copyright 2012, Elsevier (Xiong et al. 2011)
the Li–S cell showed poor capacity retention (50 % left after 50 cycles) and fast decay in terms of CE (from 97 % to below 70 % after the first two cycles) due to the severe shuttle phenomenon of polysulfides. With the addition of LiDFOB, both capacity retention and CE were improved; 2 % LiDFOB was the optimal content in this electrolyte, which yielded 70 % capacity retention after 50 cycles and 97 % CE. At 2 % LiDFOB, the a.c. impedance of the Li electrode exhibited its lowest value. The authors attributed this performance improvement to the formation of a good SEI layer on the Li anode surface that suppressed the polysulfide shuttle. The SEM images of the Li electrodes after 50 cycles also showed looser and rougher surface morphologies with a fractured surface for the Li cycled in the control electrolyte without LiDFOB additive (Fig. 4.12a). However, with 2 % LiDFOB the Li surface became relatively smooth except for some small cracks (as indicated by the rectangles in Fig. 4.12b). The XRD, XPS, and density functional theory (DFT) analyses confirmed that LiDFOB promoted the formation of a LiF-rich SEI layer on the Li metal surface, which not only protected and stabilized the Li surface but also blocked the polysulfide shuttle. Lin et al. studied the effect of P2S5 as an additive in liquid electrolytes on the performance of Li–S batteries (Lin et al. 2013b). It was found that P2S5 had two
4.2 Lithium Metal Anodes in Lithium–Sulfur Batteries
169
Fig. 4.12 SEM images of Li electrodes after 50 cycles with (a) no LiDFOB and (b) 2 % LiDFOB. Reproduced with permission—Copyright 2014, American Chemical Society (Wu et al. 2014)
functionalities: one was to promote the dissolution of Li2S in order to alleviate the capacity loss by the precipitation of Li2S, and the other was to passivate the Li anode surface so as to eliminate the polysulfide shuttle. The passivation layer was about 3–5 lm thick with granular particles less than 100 nm large, which were mainly Li3PS4. Due to the superionic conductivity of Li3PS4 for Li+, the thick but highly ionically conductive SEI layer could enhance the Li+ transportation to and from the Li anode but greatly reduce the polysulfide diffusion and reaction with the Li anode. As a result, the Li metal anode was protected; the CE and cycling performance were improved. Recently, Zu and Manthiram (2014) reported using copper acetate [Cu (OOCCH3)2 or CuAc2] as a surface stabilizer for the Li metal anode in Li–S batteries, with an electrolyte of CF3SO3Li in DOL/DME with LiNO3. The control cell (without CuAc2) had a sudden capacity decay to 0 mAh g−1 at a cycle number close to 100 cycles, likely due to the Li anode degradation. However, with the addition of 0.03 M CuAc2 in the electrolyte, the cell showed a capacity retention of 75 % at the 300th cycle, and a CE of 100 % for most of the cycles; the cell internal resistance was also lower than that of the control cell. The Li anode surfaces showed different morphologies after cycling in the two electrolytes with and without CuAc2. After the first discharge/charge cycle, the Li anode from the control electrolyte contained bulk Li2S/Li2S2 precipitates on the surface and exhibited nonuniformly deposited mossy Li (Fig. 4.13a). The Li anode from the CuAc2containing electrolyte had a smooth surface and was covered by an S- and Cu-containing passivation film (Fig. 4.13b). After 100 cycles, the Li anode from the control electrolyte showed dendritic morphology and a large amount of S-containing species (Fig. 4.13c), while the Li anode from the CuAc2-containing electrolyte was uniform (Fig. 4.13d). As shown in the cross-sectional energy dispersive spectroscopy (EDS) line scans (Fig. 4.13e, f), S was only detected in the top 20 lm of the surface film on the Li anode cycled in the CuAc2-based electrolyte, indicating the effective blocking of polysulfide penetration, thus preventing the
170
4 Application of Lithium Metal Anodes
Fig. 4.13 SEM characterizations of the Li metal surface after the first charge in the a control cell and b experimental cell and after the 100th charge in the c control cell and d experimental cell. Cross-sectional EDS line scans (for sulfur) of the Li metal after the 100th charge in the e control cell and f experimental cell. The scale bars in a–d represent 100 lm. The scales in e and f represent the position in Li metal. The cells were cycled at C/5. Reproduced with permission— Copyright 2014, American Chemical Society (Zu and Manthiram 2014)
formation of Li dendrites. The analyses by EDS, time-of-flight secondary ion mass spectrometry, XRD, and XPS demonstrated the existence of Li2S, Li2S2, CuS, Cu2S, and other electrolyte decomposition byproducts in the surface film on the Li anode from the CuAc2-containing electrolyte.
4.2 Lithium Metal Anodes in Lithium–Sulfur Batteries
171
Fig. 4.14 SEM images of the Li metal surface and cross-section of fresh Li metal (a1 and a2) and Li metal cycled in the baseline electrolyte (b1 and b2), electrolyte containing 50 % IL (c1 and c2), and electrolyte containing 75 % IL (d1 and d2) for 100 cycles at 0.2 C. Reproduced with permission—Copyright 2013, The Royal Society of Chemistry (Zheng et al. 2013b)
Zheng et al. also studied the properties and morphologies of the SEI layers formed on the Li metal surface in Li–S batteries by adding the IL 1-butyl-1-methylpyrrolidinium bis(trifluoromethanesulfonyl)-imide (Pyr14TFSI) in the electrolyte of 1 M LiTFSI/DME-DOL. It was found that the electrolyte with 75 % Pyr14TFSI in the DME-DOL-Pyr14TFSI mixture gave the Li–S batteries the best cycling stability, highest CE, and highest capacity after 30 cycles among the IL contents from 0 to 100 % in the solvent mixture. The IL-enhanced passivation film on the Li anode surface exhibited very different morphology and chemical composition, effectively protecting the Li metal from continuous attack by soluble polysulfides (see Fig. 4.14) (Zheng et al. 2013b).
4.2.2.4
Effect of Solid Electrolytes
Possible solutions to completely block Li dendrite penetration through the electrolytes include solid-state inorganic electrolytes, which normally are glassy or ceramic, and solid polymer electrolytes. In a recent review paper on Li metal anodes for rechargeable batteries, we have summarized previous achievements in using polymer electrolytes and inorganic solid-state Li-ion conductors to block the penetration of Li dendrites (Xu et al. 2014b). So far, there are several papers reporting the use of such solid-state polymer electrolytes (Marmorstein et al. 2000; Shin et al. 2002; Yu et al. 2004; Hassoun and Scrosati 2010; Liang et al. 2011a) and inorganic electrolytes (Hayashi et al. 2003, 2008; Nagao et al. 2011, 2012a, 2013; Lin et al. 2013a; Hakari et al. 2014) in Li–S related batteries. However, there are no direct reports about using solid-state polymer electrolytes and inorganic electrolytes to suppress Li dendrites in Li–S batteries. We expect that the use of solid-state
172
4 Application of Lithium Metal Anodes
inorganic electrolytes can prevent Li dendrite growth and penetration in Li–S batteries on the basis of previous reports and knowledge. However, one precaution for using such SSEs is the high interfacial resistance between the solid electrolyte with the S cathode and the Li metal anode. The ability to reduce such interfacial resistances is a big challenge in all solid-state Li–S batteries.
4.2.2.5
Effect of Ex Situ Formed Coating Layers
The protection layers on the Li anode surface are normally formed in situ in liquid electrolytes as discussed above. An alternative approach to protect the Li anode is to pre-form a thin barrier on the Li surface. Kim et al. (2013b) reported the formation of a thin Li–Al alloy layer on the Li surface to protect the Li anode. They laminated a very thin Al foil (0.8 lm) on the Li surface using pressure, and then cured at three different temperatures (25, 60 and 90 °C) for one day. The Li metal with the formed Li–Al thin layer was analyzed and tested in Li–S batteries with a high-concentration electrolyte of 3 M LiTFSI/DOL-DME. The curing temperature affected the electrochemical performance of the Li anode. The Li–Al alloy coating layer cured at 90 °C led to the lowest charge transfer resistance, the lowest polarization voltage, and the best rate capability, which were ascribed to good protection of the Li anode. However, the long-term cycling stability results showed that after 300 cycles, the discharge capacity of the Li–S cell with the Li–Al alloy layer cured at 90 °C was nearly the same as that of a pure Li anode, indicating continuous decay of this alloy protection layer. This is likely related to the high volume change of the Li–Al alloy during de-alloying/re-alloying processes that take place during the charge/discharge cycles.
4.3
Lithium Metal Anodes in Lithium-Air Batteries
Li-air batteries have much higher specific energies than most currently available primary and rechargeable batteries (Hamlen and Atwater 2001; Blurton and Sammells 1979; Arai and Hayashi 2009; Haas and Van Wesemael 2009; Visco et al. 2009; Smedley and Zhang 2009; Zhang 2009; Egashira 2009; Joerissen 2009). A significant amount of work was conducted on Li-air batteries in the 1960s and early 1970s (Gregory 1972; Blurton and Oswin 1972; Oswin 1967); however, efforts in this field decreased considerably in the 1980s because of problems associated with the stability of Li metal anodes and air electrodes, thermal management, and the reversibility of the system. Recent advances in electrode materials and electrolytes, as well as new designs for Li-air batteries, have renewed intensive efforts, especially in the development of Li-air batteries (Abraham and Jiang 1996a, b; Dobley et al. 2006c). In contrast to most other batteries that must carry both anode and cathode inside a storage system, Li-air batteries are unique in that the active cathode material (oxygen) is not stored in the battery. Instead, oxygen can be
4.3 Lithium Metal Anodes in Lithium-Air Batteries
173
absorbed from the environment and reduced by catalytic surfaces inside the air electrode. In the typical reaction in a Li-air battery using a nonaqueous electrolyte, the electrolyte will not participate in the reaction, as shown by Eq. (4.1): Li þ 1=2 O2 $ 1=2 Li2 O2
ð4:1Þ
Therefore, a very thin layered (i.e., *0.2–0.3 mm thick) carbon-based air electrode can be used to facilitate electrochemical reactions in metal-air batteries. Figure 4.15 shows the schematic of a Li-air battery based on a nonaqueous electrolyte. One important feature in this battery is that Li+ is the ion carrier transferred from the anode to the air electrode. As a result, the reaction products are accumulated in the air electrode (cathode) side rather than the anode side of the cell. Primary aqueous Li-air batteries have been used for decades in applications such as life-vest beacons; the battery is activated by sea water, and the high pH associated with LiOH formation slows Li corrosion sufficiently for operation in this application. However, water corrosion of the anode has hindered development of Li-air batteries for other practical applications (Littauer and Tsai 1977). Several approaches have been used to overcome the water corrosion problem and improve the lifetime of Li-air batteries. Abraham and Jiang (1996b) first reported a Li-air battery based on a nonaqueous electrolyte in 1996. Since then, many groups have conducted extensive work, and have documented the effects of various factors on the performance of Li-air batteries (Abraham and Jiang 1996a, b; Dobley et al. 2006b; Read et al. 2003; Kinoshita 1992; Linden and Reddy 2001; Kuboki et al. 2005; Zhang et al. 2010a; Xiao et al. 2010b; Littauer and Tsai 1977; Ye and Xu
Fig. 4.15 Schematic of reaction processes in a Li-air battery based on a nonaqueous electrolyte. Reproduced with permission —Copyright 2012, American Chemical Society (Shao et al. 2012a)
174
4 Application of Lithium Metal Anodes
2008; Kumar et al. 2010; Visco et al. 2004a, 2007; Zheng et al. 2008; Beattie et al. 2009; Abraham et al. 1997; Read 2002, 2006; Yang and Xia 2010; Dobley et al. 2006a; Shiga et al. 2008; Williford and Zhang 2009; Xiao et al. 2010a; Xu et al. 2009, 2010; Kowalczk et al. 2007; Wang and Zhou 2010; Shimonishi et al. 2010; Ogasawara et al. 2006; Debart et al. 2007a, b, 2008; Giordani et al. 2010; Laoire et al. 2009). Although a Li anode is relatively stable in a nonaqueous electrolyte, most of these Li-air batteries were investigated in a pure oxygen environment because penetration of moisture from the ambient environment will still corrode a Li electrode. To further extend the lifetime of Li-air batteries, other electrolytes have been investigated in recent years. These electrolytes include ILs (Kuboki et al. 2005; Ye and Xu 2008), SSEs (Kumar et al. 2010), and multilayer electrolytes (Visco et al. 2004a, 2007) consisting of liquid/solid, liquid/solid/solid, and liquid/solid/liquid electrolyte combinations. These multilayer electrolytes (or protected Li anodes [PLEs]) were developed by Visco et al. (2004a, 2007) and have been used successfully in various configurations, including with aqueous electrolytes. In the battery reported by Abraham and Jiang (1996a, b) and Abraham et al. (1997), the electrolyte consisted of a polymer electrolyte film composed of a mixture of polyacrylonitrile (PAN), EC, PC, and LiPF6 (in a weight ratio of 12:40:40:8, respectively) prepared inside a dry box, heated to 135 °C to obtain a homogeneous solution, and then cooled and rolled into thin membranes. The same electrolyte also was used in a carbon-based air electrode. Figure 4.16 shows the intermittent discharge curve and the open-circuit voltages of a Li|PAN-based polymer electrolyte|oxygen cell at a current density of 0.1 mA cm−2 at room temperature in an oxygen atmosphere. The cathode contained Chevron acetylene black carbon. The cell was discharged in 1.5 h increments with an open-circuit stand of about 15 min between discharges. In Fig. 4.16, the open circles represent open-circuit voltage, and the solid line represents load voltage. Abraham and Jiang also presented data suggesting that the main discharge reaction is the reduction of oxygen to form Li2O2.
Fig. 4.16 The intermittent discharge curve and the open-circuit voltages of a Li| PAN-based polymer electrolyte|oxygen cell at a current density of 0.1 mA cm−2 at room temperature in an oxygen atmosphere. Reproduced with permission—Copyright 1996, Electrochemical Society (Abraham and Jiang 1996b)
4.3 Lithium Metal Anodes in Lithium-Air Batteries
4.3.1
175
Li-Air Batteries Using Protected Lithium Electrodes
Water corrosion of the Li-metal electrode is one of the main barriers to the practical application of Li-air batteries. Therefore, protection of the Li-metal electrode is a focal point for research in this field. Solid electrolytes, such as LATP glass (Li1+x+y AlxTi2−x SiyP3−yO12 made by Ohara Inc. of Japan) have good Li-ion conductivity (*10−4 S cm−1) and are impermeable to water. This glass is stable in weak acid and alkaline electrolyte. One of the disadvantages of LATP glass is that it is not stable when in contact with Li metal. Visco et al. (2007, 2009) first solved this problem by introducing an interfacial layer (a solid layer such as Cu3N, LiPON, or nonaqueous electrolyte) between the Li metal and the Ohara glass, thus forming a protected Li electrode (PLE, Fig. 3-26b) (Visco et al. 2007, 2009). Li-air batteries using this PLE can operate in both aqueous and nonaqueous electrolytes. Figure 4. 17 shows the typical discharge curves of Li-air aqueous cells with a protected anode, employing a compliant seal, at various rates: (1) 1.0 mA cm−2; (2) 0. 5 mA cm−2; and (3) 0.2 mA cm−2. The thickness of the Li foil was *2 mm. The end of cell discharge corresponds to Li depletion. Recently, several other groups (Read and Kowalczk et al. (2007), Wang and Zhou (2010), Shimonishi et al. (2010), also used a nonaqueous or polymer electrolyte as the interfacial layer between the Li metal and the LATP glass, thereby forming a triple-electrolyte (nonaqueous electrolyte/LATP/aqueous electrolyte) structure.
4.3.2
Lithium-Air Batteries Using Solid Electrolytes
Kumar et al. (2010) reported on a solid-state Li-air battery. The Li-ion conductive solid electrolyte membrane is based on glass-ceramic (GC) and polymer-ceramic materials. This solid electrolyte is used as the ionic conductive membrane between
Fig. 4.17 Discharge of Li-air aqueous cells with a protected anode employing a compliant seal at the following discharge rates: 1 1.0 mA cm−2; 2 0.5 mA cm−2; and 3 0.2 mA cm−2. The thickness of the Li foil was *2 mm. The end of cell discharge corresponds to Li depletion. Reproduced with permission —Copyright 2009, Elsevier (Visco et al. 2009)
176
4 Application of Lithium Metal Anodes
the Li electrode and the air electrode. It also is used in the solid composite air cathode prepared from high-surface area carbon. The cell exhibited excellent thermal stability in the temperature range from 30 to 105 °C temperature range and has been tested for 40 charge-discharge cycles at current densities ranging from 0.05 to 0.25 mA cm−2. A Li-ion–conducting GC electrolyte was synthesized from the GC membrane prepared from a batch of various oxides. The composition ratio of the membrane was 18.5Li2O:6.07Al2O3:37.05GeO2:37.05P2O5 (LAGP). A schematic of the cell configuration is shown in Fig. 4.18. A Li anode and a carbon-based cathode are separated by an electrolyte laminate composed of the GC membrane with the composition 5Li2O:6.07Al2O3:37.05GeO2:37.05P2O5 (LAGP) sandwiched between two polymer composite membranes. The polymer composite adjacent to the anode is composed of LiBETI:PEO (1:8.5) with 1 wt% Li2O, whereas the polymer composite membrane adjacent to the cathode consists of LiBETI:PEO (1:8.5) with 1 wt% boron nitride. The GC electrolyte used in this work exhibits an ionic conductivity of *10−2 S cm−1 at 30 °C. The capacity of these solid-state Li-air batteries was found to increase significantly with increasing temperature. This behavior may be attributed to the interfacial resistance in the multilayer structure of Li-air batteries shown in Fig. 4.18. Although significant progress has been made on primary Li-air batteries, more work needs to be done before rechargeable Li-air batteries can be practically applied. The fundamental reaction mechanisms associated with charge and discharge of a nonaqueous Li-air cell need to be further investigated and understood. In addition, a better bifunctional catalyst needs to be developed to reduce the voltage hysteresis and improve the reversibility of the batteries, and the power rate needs to be improved. Furthermore, in the case of aqueous electrolyte-based Li-air batteries, better protective coatings for the anode are needed, as well as a means of removing the carbon dioxide before it enters the cell. For the nonaqueous Li-air
Fig. 4.18 Schematic of the Li-O2 cell and its component materials. The electrolyte laminate is composed of PC/Li2O, GC, and polymer composite (PC/boron nitride) membranes. Reproduced with permission—Copyright American Chemical Society 2010 (Kumar et al. 2010)
4.3 Lithium Metal Anodes in Lithium-Air Batteries
177
design, one of the challenges is to avoid the ingress of water, which may react with Li metal and reduce the lifetime of the batteries.
4.4
Anode-Free Lithium Batteries
A commercial Li-ion battery consists of multiple stacks of anode current collector/anode|separator|cathode/cathode current collector soaked with liquid electrolyte (such as 1 M LiPF6 in carbonate co-solvents). For example, a common Li-ion battery used in commercial electronics has a structure of Cu/graphite|separator|LiCoO2/Al or Cu/graphite|separator|LiFePO4/Al (Jiang et al. 2014; Sun et al. 2012). The operating principle of these Li-ion batteries could be described as follows. During the charging process, the Li ions are extracted out of cathode materials, diffuse through electrolyte soaked in separator, and then intercalate into the anode materials (i.e., graphite). The process is reversed during the discharging process. In this system, each component counts in the total weight and cost of the whole system, although the inactive components including separator, Cu/Al current collectors, and the packaging materials do not contribute to the usable energy. Ideally, if the Li ions extracted from a cathode can be reversibly deposited onto and stripped from a Cu current collector, then it is possible to assemble a rechargeable Li battery with a structure of Cu|separator|cathode/Al. This battery contains no active anode material as prepared, and can be called an “anode-free Li battery” (AFLB). A schematic illustration of an AFLB concept is presented in Fig. 4.19, where the state-of-the-art Li-ion battery configuration (Fig. 4.19a) is also shown as a reference. In an anode-free battery configuration, all active Li+ ions are stored inside the cathode electrode (Fig. 4.19b). During the initial charging process, Li ions are extracted from the cathode and diffuse toward the anode current collector, where they are electrodeposited as metallic Li. During the subsequent charging process, Li will be stripped from the anode and intercalated back into the
Fig. 4.19 Schematic illustration of a state-of-the-art Li-ion battery (i.e., Cu/graphite||LiFePO4/Al) and b an anode-free battery (i.e., Cu||LiFePO4/Al)
178
4 Application of Lithium Metal Anodes
cathode (Ding et al. 2013; Zheng et al. 2014; Lu et al. 2014). Traditionally, this battery structure was considered to be impossible, let alone able to survive for many cycles. In addition to the long-term concern regarding dendritic Li growth that would penetrate through the separator and short the cell (Aurbach et al. 2002; López et al. 2009; Miao et al. 2014), another direct concern with “anode-free” battery structure is that Li cycling CE is very low, usually less than 80 % in most nonaqueous electrolytes. As a consequence, the in situ plated Li supplied from the cathode electrodes will be readily consumed after fewer than ten cycles. In this regard, the key point to realize the anode-free concept is the development of a functional electrolyte in which Li deposition is dendrite free along with high CE and good cycling stability. The anode-free battery configuration exhibits several advantages. First, the absence of active anode electrode in the as-assembled batteries will reduce the weight and space reserved for the anode. In typical Li-ion batteries, the thickness of the graphite anode is similar to that of the cathode (typically a Li intercalation compound). If the graphite anode can be eliminated in anode-free batteries, the thickness of the active materials (cathode/anode) can be reduced to half of that of Li-ion batteries. Even considering other inactive materials, the energy density (Wh L−1) of the battery can be increased by more than 50 %. Second, it will save energy and cost in anode electrode preparation, including electrode slurry making, slurry coating, and drying. Last but not least, it will operate as a Li metal battery after the initial charge process, providing higher operating voltage and energy density than Li-ion batteries using graphite as the anode. Although the AFLB has many advantages, tremendous barriers need to be overcome before its practical application. The most critical barrier in these batteries is the very high CE required for an AFLB to have a meaningful lifetime. Table 4.1 shows the cycle life of a battery with different Li cycling CEs. Here we assume that the cathode of the battery has no degradation and the battery fails if its capacity drops to less than 80 % of the original values. At a CE of 99 %, the battery only has a lifetime of 22 cycles. For a rechargeable battery to have a significant lifetime, such as more than 100 cycles, the Li metal anode has to have a CE of at least 99.8 %. At a CE of larger than 99.9 %, an AFLB can have a lifetime of more than a few hundred cycles. At such a high CE, not only the Li metal anode, but all other battery components have to be stable during the electrochemical cycling process. In addition, to evaluate such a high CE, a high precision battery tester is required to
Table 4.1 CE required to retain more than 80 % of the battery capacities Coulombic efficiency (%)
Cycles to maintain a capacity >80 % of initial value
99 99.7 99.8 99.9 99.99
22 74 112 223 2231
4.4 Anode-Free Lithium Batteries
179
obtain a reliable evaluation of the Li cycling CE. Further development of this technology may lead to practical application of these batteries in the future. Using solid-state electrolytes such as LiPON has been regarded as a promising solution for anode-free batteries, owing to their good compatibility with Li (Neudecker et al. 2000; Bates et al. 1993, 1995; Lee et al. 2004). However, the intrinsically low conductivity of LiPON and their high cost has limited the application of these batteries (Wang et al. 2015a; Bates et al. 1993, 1995). Therefore, only thin-film electrodes with an areal capacity of *100 lAh cm−2 have been reported, limiting their applications in microelectronic devices that do not require high energy or power supply (Neudecker et al. 2000). A liquid organic electrolyte with high conductivity and wettability can be used for the thick electrode in Li metal cells. However, the traditional carbonate-based electrolyte system has poor compatibility with Li metal (due to Li dendrite growth and low CE) and therefore could not be used for anode-free cell design (Woo et al. 2014; Aurbach et al. 2002; Xu et al. 2014b). Qian et al. (2015b) demonstrated that the highly concentrated electrolytes composed of ether solvents and the LiFSI salt enabled high-rate cycling of a Li metal anode with high CE (up to 99.1 %) without dendrite growth; this was ascribed to the enhanced solvent coordination and improved availability of Li+ ion concentration in the electrolyte. They also demonstrated an anode-free Li cell design (Cu||LiFePO4) with highly concentrated electrolyte (4 M LiFSI/DME) and delivered high initial discharge capacities close to the nominal capacity of the cathode. The battery retained *60 % of its original capacity after 50 cycles along with an average CE >99 % during cycling. The CE of these batteries can be further improved by adopting a slow charge/fast discharge protocol resulting in an exceptionally high CE of more than 99.8 % (Qian et al. 2016). The in situ formed Li (during charging) is largely stabilized due to minimized reactions between the plated Li and concentrated electrolyte. These works and the insights obtained for the SEI characteristics demonstrate a possible path forward for enabling Li metal as a highly reversible and safe practical battery anode.
References Abraham KM, Jiang Z (1996a) Solid polymer electrolyte-based oxygen batteries. US 5510209 Abraham KM, Jiang Z (1996b) A polymer electrolyte-based rechargeable lithium/oxygen battery. J Electrochem Soc 143(1):1–5 Abraham KM, Jiang Z, Carroll B (1997) Highly conductive PEO-like polymer electrolytes. Chem Mater 9(9):1978–1988 Agostini M, Aihara Y, Yamada T, Scrosati B, Hassoun J (2013) A lithium–sulfur battery using a solid, glass-type P2S5–Li2S electrolyte. Solid State Ionics 244:48–51. doi:http://dx.doi.org/10. 1016/j.ssi.2013.04.024 Arai H, Hayashi M (2009) Secondary batteries-metal-air systems: overview (secondary and primary). In: Garche J, Dyer C, Moseley P, Ogumi Z, Rand D, Scrosati B (eds) Encyclopedia of electrochemical power sources, vol 5. Elsevier, Amsterdam, pp 347–355
180
4 Application of Lithium Metal Anodes
Aurbach D, Zinigrad E, Teller H, Dan P (2000) Factors which limit the cycle life of rechargeable lithium (metal) batteries. J Electrochem Soc 147:1274–1279 Aurbach D, Zinigrad E, Cohen Y, Teller H (2002) A short review of failure mechanisms of lithium metal and lithiated graphite anodes in liquid electrolyte solutions. Solid State Ionics 148:405– 416 Aurbach D, Pollak E, Elazari R, Salitra G, Kelley CS, Affinito J (2009) On the surface chemical aspects of very high energy density, rechargeable Li–sulfur batteries. J Electrochem Soc 156 (8):A694–A702. doi:10.1149/1.3148721 Autolib’ Bluecar carsharing service, Paris, France, by Mariordo (Mario Roberto Durán Ortiz) is licensed under CC BY-SA 3.0 and found on Wikipedia https://en.wikipedia.org/wiki/Bollor% C3%A9_Bluecar#/media/File:Paris_Autolib_06_2012_Bluecar_2907.JPG. Accessed 31 May 2016 Barchasz C, Molton F, Duboc C, Leprêtre J-C, Patoux S, Alloin F (2012) Lithium/sulfur cell discharge mechanism: an original approach for intermediate species identification. Anal Chem 84(9):3973–3980. doi:10.1021/ac2032244 Barchasz C, Lepretre JC, Patoux S, Alloin F (2013a) Revisiting TEGDME/DIOX binary electrolytes for lithium/sulfur batteries: importance of solvation ability and additives. J Electrochem Soc 160(3):A430–A436. doi:10.1149/2.022303jes Barchasz C, Leprêtre J-C, Patoux S, Alloin F (2013b) Electrochemical properties of ether-based electrolytes for lithium/sulfur rechargeable batteries. Electrochim Acta 89:737–743. doi:10. 1016/j.electacta.2012.11.001 Barghamadi M, Kapoor A, Wen C (2013) A review on Li-S batteries as a high efficiency rechargeable lithium battery. J Electrochem Soc 160(8):A1256–A1263. doi:10.1149/2. 096308jes Bates JB, Dudney NJ, Gruzalski GR, Zuhr RA, Choudhury A, Luck CF, Robertson JD (1993) Fabrication and characterization of amorphous lithium electrolyte thin-films and rechargeable thin-film batteries. J Power Sources 43(1–3):103–110. doi:10.1016/0378-7753(93)80106-y Bates JB, Dudney NJ, Lubben DC, Gruzalski GR, Kwak BS, Yu X, Zuhr RA (1995) Thin-film rechargeable lithium batteries. J Power Sources 54(1):58–62. doi:http://dx.doi.org/10.1016/ 0378-7753(94)02040-A Beattie SD, Manolescu DM, Blair SL (2009) High-capacity lithium-air cathodes. J Electrochem Soc 156(1):A44–A47. doi:10.1149/1.3005989 Blurton KF, Oswin HG (1972) Refuelable batteries. In: Proceedings of the symposium on non-fossil chemical fuels, preprints of papers presented at the 163rd national meeting of the American Chemical Society, vol 2, Boston, Massachusetts, 10–14 April 1972, American Chemical Society, Division of Fuel Chemistry, Washington DC, pp 48–69 Blurton KF, Sammells AF (1979) Metal-air batteries—their status and potential—review. J Power Sources 4(4):263–279 Bresser D, Passerini S, Scrosati B (2013) Recent progress and remaining challenges in sulfur-based lithium secondary batteries—a review. Chem Commun 49(90):10545–10562. doi:10.1039/C3CC46131A Bruce PG, Freunberger SA, Hardwick LJ, Tarascon J-M (2012) Li-O2 and Li-S batteries with high energy storage. Nat Mater 11(1):19–29 Brückner J, Thieme S, Böttger-Hiller F, Bauer I, Grossmann HT, Strubel P, Althues H, Spange S, Kaskel S (2014) Carbon-based anodes for lithium sulfur full cells with high cycle stability. Adv Funct Mater 24(9):1284–1289. doi:10.1002/adfm.201302169 Bryngelsson H, Stjerndahl M, Gustafsson T, Edström K (2007) How dynamic is the SEI? J Power Sources 174(2):970–975. doi:10.1016/j.jpowsour.2007.06.050 Bucur CB, Muldoon J, Lita A, Schlenoff JB, Ghostine RA, Dietz S, Allred G (2013) Ultrathin tunable ion conducting nanomembranes for encapsulation of sulfur cathodes. Energy Environ Sci 6(11):3286–3290. doi:10.1039/C3EE42739K Busche MR, Adelhelm P, Sommer H, Schneider H, Leitner K, Janek J (2014) Systematical electrochemical study on the parasitic shuttle-effect in lithium-sulfur-cells at different
References
181
temperatures and different rates. J Power Sources 259:289–299. doi:10.1016/j.jpowsour.2014. 02.075 Chen L, Shaw LL (2014) Recent advances in lithium–sulfur batteries. J Power Sources 267:770– 783. doi:10.1016/j.jpowsour.2014.05.111 Chung S-H, Manthiram A (2014a) Carbonized eggshell membrane as a natural polysulfide reservoir for highly reversible Li-S batteries. Adv Mater 26(9):1360–1365. doi:10.1002/adma. 201304365 Chung S-H, Manthiram A (2014b) A natural carbonized leaf as polysulfide diffusion inhibitor for high-performance lithium-sulfur battery cells. ChemSusChem 7(6):1655–1661. doi:10.1002/ cssc.201301287 Chung S-H, Manthiram A (2014c) High-performance Li–S batteries with an ultra-lightweight MWCNT-coated separator. J Phys Chem Lett 5(11):1978–1983. doi:10.1021/jz5006913 Chung SH, Manthiram A (2014d) Bifunctional separator with a light-weight carbon-coating for dynamically and statically stable lithium-sulfur batteries. Adv Funct Mater 24(33):5299–5306. doi:10.1002/adfm.201400845 Debart A, Bao J, Armstrong G, Bruce PG (2007a) Effect of catalyst on the performance of rechargeable lithium/air batteries. ECS Trans 3(27):225–232. doi:10.1149/1.2793594 Debart A, Bao J, Armstrong G, Bruce PG (2007b) An O2 cathode for rechargeable lithium batteries: the effect of a catalyst. J Power Sources 147(2):1177–1182 Debart A, Paterson AJ, Bao J, Bruce PG (2008) Alpha-MnO2 nanowires: a catalyst for the O-2 electrode in rechargeable lithium batteries. Angew Chem Int Edit 47(24):4521–4524. doi:10. 1002/anie.200705648 Demir-Cakan R, Morcrette M, Gangulibabu GA, Dedryvere R, Tarascon J-M (2013) Li-S batteries: simple approaches for superior performance. Energy Environ Sci 6(1):176–182. doi:10.1039/C2EE23411D Ding F, Xu W, Graff GL, Zhang J, Sushko ML, Chen X, Shao Y, Engelhard MH, Nie Z, Xiao J, Liu X, Sushko PV, Liu J, Zhang J-G (2013) Dendrite-free lithium deposition via self-healing electrostatic shield mechanism. J Am Chem Soc 135(11):4450–4456. doi:10.1021/ja312241y Dobley A, Morein C, Abraham KM (2006a) Cathode optimization for lithium-air batteries. Paper presented at the 208th meeting of the electrochemical society abstracts: energy technology and battery joint general session, 16–21 Oct 2005, Los Angeles, California Dobley A, Morein C, Roark R, Abraham KM (2006b) Paper presented at the proceedings of the 42nd power sources conference, 12–15 June 2006, Philadelphia, Pennsylvania Dobley A, Morein C, Roark R, Abraham KM (2006c) Large prototype lithium air batteries. In: Proceedings of the 42nd power sources conference, Philadelphia, Pennsylvania, 12–15 June 2006. U.S. Army Communications-Electronics Command, Fort Monmouth, New Jersey Egashira M (2009) Secondary batteries-metal-air systems: iron-air (secondary and primary). In: Garche J, Dyer C, Moseley P, Ogumi Z, Rand D, Scrosati B (eds) Encyclopedia of electrochemical power sources, vol 5. Elsevier, Amsterdam, pp 372–375 Elazari R, Salitra G, Gershinsky G, Garsuch A, Panchenko A, Aurbach D (2012) Rechargeable lithiated silicon–sulfur (SLS) battery prototypes. Electrochem Commun 14(1):21–24. doi:10. 1016/j.elecom.2011.10.020 Evers S, Nazar LF (2012) New approaches for high energy density lithium-sulfur battery cathodes. Acc Chem Res 46(5):1135–1143. doi:10.1021/ar3001348 Fu Y, Su Y-S, Manthiram A (2013) Highly reversible lithium/dissolved polysulfide batteries with carbon nanotube electrodes. Angew Chem Int Ed 52(27):6930–6935. doi:10.1002/anie. 201301250 Gao J, Lowe MA, Kiya Y, Abruña HD (2011) Effects of liquid electrolytes on the charge-discharge performance of rechargeable lithium/sulfur batteries: electrochemical and in-situ X-ray absorption spectroscopic studies. J Phys Chem C 115(50):25132–25137. doi:10. 1021/jp207714c Giordani V, Freunberger SA, Bruce PG, Tarascon JM, Larcher D (2010) H2O2 decomposition reaction as selecting tool for catalysts in Li-O-2 cells. Electrochem Solid State Lett 13(12): A180–A183. doi:10.1149/1.3494045
182
4 Application of Lithium Metal Anodes
Girishkumar G, McCloskey B, Luntz AC, Swanson S, Wilcke W (2010) Lithium−air battery: promise and challenges. J Phys Chem Lett 1:2193–2203 Gregory DP (1972) Metal-air batteries. Mills and Boon, London Haas O, Van Wesemael J (2009) Secondary batteries-metal-air systems: zinc-air: electrical recharge. In: Garche J, Dyer C, Moseley P, Ogumi Z, Rand D, Scrosati B (eds) Encyclopedia of electrochemical power sources, vol 5. Elsevier, Amsterdam, pp 384–392 Hakari T, Nagao M, Hayashi A, Tatsumisago M (2014) Preparation of composite electrode with Li2S–P2S5 glasses as active materials for all-solid-state lithium secondary batteries. Solid State Ionics 262:147–150. doi:10.1016/j.ssi.2013.09.023 Hamlen RP, Atwater TB (2001) Metal/air batteries. In: Linden D, Reddy T (eds) Handbook of batteries, 3rd edn. McGraw Hill, New York, pp 38.31–38.53 Hassoun J, Scrosati B (2010) A high-performance polymer tin sulfur lithium ion battery. Angew Chem Int Ed 49(13):2371–2374. doi:10.1002/anie.200907324 Hayashi A, Ohtomo T, Mizuno F, Tadanaga K, Tatsumisago M (2003) All-solid-state Li/S batteries with highly conductive glass–ceramic electrolytes. Electrochem Commun 5(8):701– 705. doi:10.1016/S1388-2481(03)00167-X Hayashi A, Ohtsubo R, Ohtomo T, Mizuno F, Tatsumisago M (2008) All-solid-state rechargeable lithium batteries with Li2S as a positive electrode material. J Power Sources 183(1):422–426. doi:10.1016/j.jpowsour.2008.05.031 Heine J, Krüger S, Hartnig C, Wietelmann U, Winter M, Bieker P (2014) Coated lithium powder (CLiP) electrodes for lithium-metal batteries. Adv Energy Mater 4(5). doi:10.1002/aenm. 201300815 Huang J-Q, Zhang Q, Zhang S-M, Liu X-F, Zhu W, Qian W-Z, Wei F (2013) Aligned sulfur-coated carbon nanotubes with a polyethylene glycol barrier at one end for use as a high efficiency sulfur cathode. Carbon 58:99–106. doi:10.1016/j.carbon.2013.02.037 Huang C, Xiao J, Shao Y, Zheng J, Bennett WD, Lu D, Saraf LV, Engelhard M, Ji L, Zhang J, Li X, Graff GL, Liu J (2014a) Manipulating surface reactions in lithium-sulphur batteries using hybrid anode structures. Nat Commun 5:3015. doi:10.1038/ncomms4015 Huang J-Q, Zhang Q, Peng H-J, Liu X-Y, Qian W-Z, Wei F (2014b) Ionic shield for polysulfides towards highly-stable lithium-sulfur batteries. Energy Environ Sci 7(1):347–353. doi:10.1039/ C3EE42223B Jeddi K, Zhao Y, Zhang Y, Konarov A, Chen P (2013) Fabrication and characterization of an effective polymer nanocomposite electrolyte membrane for high performance lithium/sulfur batteries. J Electrochem Soc 160(8):A1052–A1060. doi:10.1149/2.010308jes Jeddi K, Sarikhani K, Qazvini NT, Chen P (2014) Stabilizing lithium/sulfur batteries by a composite polymer electrolyte containing mesoporous silica particles. J Power Sources 245:656–662. doi:10.1016/j.jpowsour.2013.06.147 Jeong T-G, Moon YH, Chun H-H, Kim HS, Cho BW, Kim Y-T (2013) Free standing acetylene black mesh to capture dissolved polysulfide in lithium sulfur batteries. Chem Commun 49 (94):11107–11109. doi:10.1039/C3CC46358C Ji XL, Lee KT, Nazar LF (2009) Nature Mater 8:500–506 Jiang J, Shi W, Zheng J, Zuo P, Xiao J, Chen X, Xu W, Zhang J-G (2014) Optimized operating range for large-format LiFePO4/graphite batteries. J Electrochem Soc 161(3):A336–A341. doi:10.1149/2.052403jes Jin Z, Xie K, Hong X (2013a) Electrochemical performance of lithium/sulfur batteries using perfluorinated ionomer electrolyte with lithium sulfonyl dicyanomethide functional groups as functional separator. RSC Adv 3(23):8889–8898. doi:10.1039/C3RA41517A Jin Z, Xie K, Hong X, Hu Z (2013b) Capacity fading mechanism in lithium sulfur cells using poly (ethylene glycol)-borate ester as plasticizer for polymer electrolytes. J Power Sources 242:478– 485. doi:10.1016/j.jpowsour.2013.05.086 Joerissen L (2009) Secondary batteries-metal-air systems: bifunctional oxygen electrodes. In: Garche J, Dyer C, Moseley P, Ogumi Z, Rand D, Scrosati B (eds) Encyclopedia of electrochemical power sources, vol 5. Elsevier, Amsterdam, pp 356–371
References
183
Kim H, Lee JT, Yushin G (2013a) High temperature stabilization of lithium–sulfur cells with carbon nanotube current collector. J Power Sources 226:256–265. doi:10.1016/j.jpowsour. 2012.10.028 Kim H, Lee JT, Lee D-C, Oschatz M, Cho WI, Kaskel S, Yushin G (2013b) Enhancing performance of Li–S cells using a Li–Al alloy anode coating. Electrochem Commun 36:38–41. doi:10.1016/j.elecom.2013.09.002 Kim H, Wu F, Lee JT, Nitta N, Lin H-T, Oschatz M, Cho WI, Kaskel S, Borodin O, Yushin G (2014) In situ formation of protective coatings on sulfur cathodes in lithium batteries with LiFSI-based organic electrolytes. Adv Energy Mater. doi:10.1002/aenm.201401792 Kinoshita K (1992) Electrochemical oxygen technology. The electochemical society series. John Wiley & Sons, New York Kinoshita S, Okuda K, Machida N, Shigematsu T (2014) Additive effect of ionic liquids on the electrochemical property of a sulfur composite electrode for all-solid-state lithium–sulfur battery. J Power Sources 269:727–734. doi:10.1016/j.jpowsour.2014.07.055 Kowalczk I, Read J, Salomon M (2007) Li-air batteries: a classic example of limitations owing to solubilities. Pure Appl Chem 79(5):851–860. doi:10.1351/pac200779050851 Kuboki T, Okuyama T, Ohsaki T, Takami N (2005) Lithium-air batteries using hydrophobic room temperature ionic liquid electrolyte. J Power Sources 146(1–2):766–769. doi:10.1016/j. jpowsour.2005.03.082 Kulisch J, Sommer H, Brezesinski T, Janek J (2014) Simple cathode design for Li-S batteries: cell performance and mechanistic insights by in operando X-ray diffraction. Phys Chem Chem Phys 16(35):18765–18771. doi:10.1039/C4CP02220C Kumar B, Kumar J, Leese R, Fellner JP, Rodrigues SJ, Abraham KM (2010) A solid-state, rechargeable, long cycle life lithium-air battery. J Electrochem Soc 157(1):A50–A54. doi:10. 1149/1.3256129 Laoire CO, Mukerjee S, Abraham KM, Plichta EJ, Hendrickson MA (2009) Elucidating the mechanism of oxygen reduction for lithium-air battery applications. J Phys Chem C 113 (46):20127–20134. doi:10.1021/jp908090s Lee YM, Choi N-S, Park JH, Park J-K (2003) Electrochemical performance of lithium/sulfur batteries with protected Li anodes. J Power Sources 119–121:964–972. doi:10.1016/s03787753(03)00300-8 Lee S-H, Tracy E, Liu P (2004) Buried anode lithium thin film battery and process for forming the same. USA Patent 6,805,999, 19 October 2004 Lee J-S, Kim ST, Cao R, Choi N-S, Liu M, Lee KT, Cho J (2011) Adv Energy Mater 1:34–50 Li W, Hicks-Garner J, Wang J, Liu J, Gross AF, Sherman E, Graetz J, Vajo JJ, Liu P (2014) V2O5 polysulfide anion barrier for long-lived Li-S batteries. Chem Mater 26(11):3404–3410 Liang X, Wen Z, Liu Y, Zhang H, Huang L, Jin J (2011a) Highly dispersed sulfur in ordered mesoporous carbon sphere as a composite cathode for rechargeable polymer Li/S battery. J Power Sources 196(7):3655–3658. doi:10.1016/j.jpowsour.2010.12.052 Liang X, Wen Z, Liu Y, Wu M, Jin J, Zhang H, Wu X (2011b) Improved cycling performances of lithium sulfur batteries with LiNO3-modified electrolyte. J Power Sources 196(22):9839–9843. doi:10.1016/j.jpowsour.2011.08.027 Lin Z, Liu Z, Dudney NJ, Liang C (2013a) Lithium superionic sulfide cathode for all-solid lithium-sulfur batteries. ACS Nano 7(3):2829–2833. doi:10.1021/nn400391h Lin Z, Liu Z, Fu W, Dudney NJ, Liang C (2013b) Phosphorous pentasulfide as a novel additive for high-performance lithium-sulfur batteries. Adv Funct Mater 23(8):1064–1069. doi:10.1002/ adfm.201200696 Linden D, Reddy T (2001) Handbook of batteries, 3rd edn. McGraw-Hill, New York Littauer EL, Tsai KC (1977) Corrosion of lithium in alkaline-solution. J Electrochem Soc 124 (6):850–855 Liu C, Ma X, Xu F, Zheng L, Zhang H, Feng W, Huang X, Armand M, Nie J, Chen H, Zhou Z (2014a) Ionic liquid electrolyte of lithium bis(fluorosulfonyl)imide/N-MethylN-propylpiperidinium bis(fluorosulfonyl)imide for Li/natural graphite cells: effect of
184
4 Application of Lithium Metal Anodes
concentration of lithium salt on the physicochemical and electrochemical properties. Electrochim Acta 149:370–385. doi:10.1016/j.electacta.2014.10.048 Liu Z, Zhang X-H, Lee C-S (2014b) A stable high performance Li-S battery with a polysulfide ion blocking layer. J Mater Chem A 2(16):5602–5605. doi:10.1039/C4TA00015C López CM, Vaughey JT, Dees DW (2009) Morphological transitions on lithium metal anodes. J Electrochem Soc 156(9):A726–A729. doi:10.1149/1.3158548 Lu Y, Tu Z, Archer LA (2014) Stable lithium electrodeposition in liquid and nanoporous solid electrolytes. Nat Mater 13(10):961–969. doi:10.1038/nmat4041 Lv D, Shao Y, Lozano T, Bennett WD, Graff GL, Polzin B, Zhang J-G, Engelhard MH, Saenz NT, Henderson WA, Bhattacharya P, Liu J, Xiao J (2015) Failure mechanism for fast-charged lithium metal batteries with liquid electrolytes. Adv Energy Mater 5(3):1400993. doi:10.1002/ aenm.201400993 Ma G, Wen Z, Jin J, Wu M, Wu X, Zhang J (2014) Enhanced cycle performance of Li–S battery with a polypyrrole functional interlayer. J Power Sources 267:542–546. doi:10.1016/j. jpowsour.2014.05.057 Manthiram A, Fu Y, Su Y-S (2012) Challenges and prospects of lithium-sulfur batteries. Acc Chem Res 46(5):1125–1134. doi:10.1021/ar300179v Manthiram A, Fu Y, Chung SH, Zu C, Su YS (2014) Rechargeable lithium-sulfur batteries. Chem Rev 114(23):11751–11787. doi:10.1021/cr500062v Marmorstein D, Yu TH, Striebel KA, McLarnon FR, Hou J, Cairns EJ (2000) Electrochemical performance of lithium/sulfur cells with three different polymer electrolytes. J Power Sources 89(2):219–226. doi:10.1016/S0378-7753(00)00432-8 Miao R, Yang J, Feng X, Jia H, Wang J, Nuli Y (2014) Novel dual-salts electrolyte solution for dendrite-free lithium-metal based rechargeable batteries with high cycle reversibility. J Power Sources 271:291–297. doi:10.1016/j.jpowsour.2014.08.011 Mikhaylik YV (2008) Electrolytes for lithium sulfur cells. USA Patent 7,354,680, 8 April 2008 Munichandraiah N, Scanlon LG, Marsh RA (1998) Surface films of lithium: an overview of electrochemical studies. J Power Sources 72:203–210 Nagao M, Hayashi A, Tatsumisago M (2011) Sulfur–carbon composite electrode for all-solid-state Li/S battery with Li2S–P2S5 solid electrolyte. Electrochim Acta 56(17):6055–6059. doi:10. 1016/j.electacta.2011.04.084 Nagao M, Hayashi A, Tatsumisago M (2012a) High-capacity Li2S-nanocarbon composite electrode for all-solid-state rechargeable lithium batteries. J Mater Chem 22(19):10015–10020. doi:10.1039/C2JM16802B Nagao M, Hayashi A, Tatsumisago M (2012b) Fabrication of favorable interface between sulfide solid electrolyte and Li metal electrode for bulk-type solid-state Li/S battery. Electrochem Commun 22:177–180. doi:10.1016/j.elecom.2012.06.015 Nagao M, Imade Y, Narisawa H, Kobayashi T, Watanabe R, Yokoi T, Tatsumi T, Kanno R (2013) All-solid-state Li–sulfur batteries with mesoporous electrode and thio-LISICON solid electrolyte. J Power Sources 222:237–242. doi:10.1016/j.jpowsour.2012.08.041 Nazar LF, Cuisinier M, Pang Q (2014) Lithium-sulfur batteries. MRS Bull 39(05):436–442. doi:10.1557/mrs.2014.86 Neudecker BJ, Dudney NJ, Bates JB (2000) “Lithium-free” thin-film battery with in situ plated Li anode. J Electrochem Soc 147(2):517–523 Ogasawara T, Debart A, Holzapfel M, Novak P, Bruce PG (2006) Rechargeable Li2O2 electrode for lithium batteries. J Am Chem Soc 128(4):1390–1393. doi:10.1021/ja056811q Oh S, Yoon W (2014) Effect of polypyrrole coating on Li powder anode for lithium-sulfur secondary batteries. Int J Precis Eng Manuf 15(7):1453–1457. doi:10.1007/s12541-014-0490-y Oh SJ, Lee JK, Yoon WY (2014) Preventing the dissolution of lithium polysulfides in lithium-sulfur cells by using Nafion-coated cathodes. ChemSusChem 7(9):2562–2566. doi:10. 1002/cssc.201402318 Oswin HG (1967) Performance forecast of selected static energy conversion devices. In: Sherman GW, Devol L (eds) Proceedings of the 29th meeting of AGARD propulsion and
References
185
energetics panel, Liege, Belgium, 12–16 June 1967. Air Force Aero Proplusion Laboratory, Wright-Patterson Air Force Base, Ohio, p 397 Park J-W, Ueno K, Tachikawa N, Dokko K, Watanabe M (2013) Ionic liquid electrolytes for lithium-sulfur batteries. J Phys Chem C 117(40):20531–20541. doi:10.1021/jp408037e Qian J, Henderson WA, Xu W, Bhattacharya P, Engelhard M, Borodin O, Zhang JG (2015) High rate and stable cycling of lithium metal anode. Nat Commun 6:6362. doi:10.1038/ ncomms7362 Qian JF, Adams BD, Zheng JM, Xu W, Henderson WA, Wang J, Bowden ME, Xu SC, Hu JZ, Zhang JZ (2016) Anode-free rechargeable lithium metal batteries. Adv Funct Mater 2016. doi:10.1002/adfm.201602353 Ratnakumar BV, Smart MC, Surampudi S (2001) Effects of SEI on the kinetics of lithium intercalation. J Power Sources 97–98:137–139. doi:10.1016/S0378-7753(01)00682-6 Read J (2002) Characterization of the lithium/oxygen organic electrolyte battery. J Electrochem Soc 149(9):A1190–A1195. doi:10.1149/1.1498256 Read J (2006) Ether-based electrolytes for the lithium/oxygen organic electrolyte battery. J Electrochem Soc 153(1):A96–A100. doi:10.1149/1.2131827 Read J, Mutolo K, Ervin M, Behl W, Wolfenstine J, Driedger A, Foster D (2003) Oxygen transport properties of organic electrolytes and performance of lithium/oxygen battery. J Electrochem Soc 150(10):A1351–A1356. doi:10.1149/1.1606454 Scheers J, Fantini S, Johansson P (2014) A review of electrolytes for lithium-sulphur batteries. J Power Sources 255:204–218. doi:10.1016/j.jpowsour.2014.01.023 Shao Y, Park S, Xiao J, Zhang J-G, Wang Y, Liu AJ (2012a) Electrocatalysts for nonaqueous lithium-air batteries: status. Challenges, and Perspective. ACS Catalysis 845 Shiga T, Nakano H, Imagawa H (2008) Non-aqueous air battery and catalyst therefor. US Patent Application 2008/0299456 Shimonishi Y, Zhang T, Johnson P, Imanishi N, Hirano A, Takeda Y, Yamamoto O, Sammes N (2010) A study on lithium/air secondary batteries-Stability of NASICON-type glass ceramics in acid solutions. J Power Sources 195(18):6187–6191. doi:10.1016/j.jpowsour.2009.11.023 Shin JH, Kim KW, Ahn HJ, Ahn JH (2002) Electrochemical properties and interfacial stability of (PEO)10LiCF3SO3–TinO2n − 1 composite polymer electrolytes for lithium/sulfur battery. Mater Sci Eng, B 95(2):148–156. doi:10.1016/S0921-5107(02)00226-X Smedley S, Zhang XG (2009) Secondary batteries-metal-air systems: zinc-air: hydraulic recharge. In: Garche J, Dyer C, Moseley P, Ogumi Z, Rand D, Scrosati B (eds) Encyclopedia of electrochemical power sources, vol 5. Elsevier, Amsterdam, pp 393–403 Song JH, Yeon JT, Jang JY, Han JG, Lee SM, Choi NS (2013a) Effect of fluoroethylene carbonate on electrochemical performances of lithium electrodes and lithium-sulfur batteries. J Electrochem Soc 160(6):A873–A881. doi:10.1149/2.101306jes Song MK, Cairns EJ, Zhang Y (2013b) Lithium/sulfur batteries with high specific energy: old challenges and new opportunities. Nanoscale 5(6):2186–2204. doi:10.1039/c2nr33044j Su Y-S, Manthiram A (2012) Lithium–sulphur batteries with a microporous carbon paper as a bifunctional interlayer. Nat Commun 3:1166 Sun Y-K, Chen Z, Noh H-J, Lee D-J, Jung H-G, Ren Y, Wang S, Yoon CS, Myung S-T, Amine K (2012) Nanostructured high-energy cathode materials for advanced lithium batteries. Nat Mater 11(11):942–947. doi:10.1038/nmat3435 Suo L, Hu YS, Li H, Armand M, Chen L (2013) A new class of solvent-in-salt electrolyte for high-energy rechargeable metallic lithium batteries. Nat Commun 4:1481. doi:10.1038/ ncomms2513 Tang Q, Shan Z, Wang L, Qin X, Zhu K, Tian J, Liu X (2014) Nafion coated sulfur–carbon electrode for high performance lithium–sulfur batteries. J Power Sources 246:253–259. doi:10. 1016/j.jpowsour.2013.07.076 Tobishima S, Sakurai Y, Yamaki J (1996) Safety characteristics of rechargeable lithium metal cells. In: 8th international meeting on lithium batteries, p 362. doi:10.1016/S0378-7753(96) 02584-0
186
4 Application of Lithium Metal Anodes
Unemoto A, Ogawa H, Gambe Y, Honma I (2014) Development of lithium-sulfur batteries using room temperature ionic liquid-based quasi-solid-state electrolytes. Electrochim Acta 125:386– 394. doi:10.1016/j.electacta.2014.01.105 Vaughey JT, Liu G, Zhang J-G (2014) Stabilizing the surface of lithium metal. MRS Bull 39 (05):429–435. doi:10.1557/mrs.2014.88 Visco SJ, Nimon E, De Jonghe LC, Katz B, Chu MY (2004a) Lithium fuel cells. Paper presented at the proceedings of the 12th international meeting on lithium batteries, 27 June–2 July 2004, Nara, Japan Visco SJ, Katz B, Nimon YS, De Jonghe LC (2007) Protected active metal electrode and battery cell structures with non-aqueous interlayer architecture. US 7282295 Visco SJ, Nimon E, De Jonghe C (2009) Secondary batteries-metal-air systems: lithium-air. In: Garche J, Dyer C, Moseley P, Ogumi Z, Rand D, Scrosati B (eds) Encyclopedia of electrochemical power sources, vol 5. Elsevier, Amsterdam, pp 376–383 Wang YG, Zhou HS (2010) A lithium-air battery with a potential to continuously reduce O-2 from air for delivering energy. J Power Sources 195(1):358–361. doi:10.1016/j.jpowsour.2009.06. 109 Wang X, Wang Z, Chen L (2013) Reduced graphene oxide film as a shuttle-inhibiting interlayer in a lithium–sulfur battery. J Power Sources 242:65–69. doi:10.1016/j.jpowsour.2013.05.063 Wang J, Lin F, Jia H, Yang J, Monroe CW, NuLi Y (2014) Towards a safe lithium-sulfur battery with a flame-inhibiting electrolyte and a sulfur-based composite cathode. Angew Chem Int Ed 53(38):10099–10104. doi:10.1002/anie.201405157 Wang D, Zhong G, Li Y, Gong Z, McDonald MJ, Mi J-X, Fu R, Shi Z, Yang Y (2015) Enhanced ionic conductivity of Li3.5Si0.5P0.5O4 with addition of lithium borate. Solid State Ionics 283:109–114. doi:10.1016/j.ssi.2015.10.009 Whittingham MS (2004) Lithium batteries and cathode materials. Chem Rev 104(10):4271–4302. doi:10.1021/cr020731c Whittingham MS (2012) History, evolution, and future status of energy storage. In: Proceedings of IEE 100 (special centennial issue), 1518–1534. doi:10.1109/JPROC.2012.2190170 Williford RE, Zhang JG (2009) Air electrode design for sustained high power operation of Li/air batteries. J Power Sources 194(2):1164–1170. doi:10.1016/j.jpowsour.2009.06.005 Woo J-J, Maroni VA, Liu G, Vaughey JT, Gosztola DJ, Amine K, Zhang Z (2014) Symmetrical impedance study on inactivation induced degradation of lithium electrodes for batteries beyond lithium-ion. J Electrochem Soc 161(5):A827–A830. doi:10.1149/2.089405jes Wu M, Wen Z, Jin J, Cui Y (2013) Effects of combinatorial AlCl3 and pyrrole on the SEI formation and electrochemical performance of Li electrode. Electrochim Acta 103:199–205. doi:10.1016/j.electacta.2013.03.181 Wu F, Qian J, Chen R, Lu J, Li L, Wu H, Chen J, Zhao T, Ye Y, Amine K (2014) An effective approach to protect lithium anode and improve cycle performance for Li-S batteries. ACS Appl Mater Interfaces 6(17):15542–15549. doi:10.1021/am504345s Xiao J, Xu W, Wang DY, Zhang JG (2010a) Hybrid air-electrode for Li/air batteries. J Electrochem Soc 157(3):A294–A297. doi:10.1149/1.3280281 Xiao J, Wang DH, Xu W, Wang DY, Williford RE, Liu J, Zhang JG (2010b) Optimization of air electrode for Li/air batteries. J Electrochem Soc 157(4):A487–A492. doi:10.1149/1.3314375 Xiong S, Kai X, Hong X, Diao Y (2011) Effect of LiBOB as additive on electrochemical properties of lithium-sulfur batteries. Ionics 18(3):249–254. doi:10.1007/s11581-011-0628-1 Xiong S, Xie K, Diao Y, Hong X (2012) Properties of surface film on lithium anode with LiNO3 as lithium salt in electrolyte solution for lithium-sulfur batteries. Electrochim Acta 83:78–86. doi:10.1016/j.electacta.2012.07.118 Xiong S, Xie K, Diao Y, Hong X (2013) On the role of polysulfides for a stable solid electrolyte interphase on the lithium anode cycled in lithium-sulfur batteries. J Power Sources 236:181– 187. doi:10.1016/j.jpowsour.2013.02.072 Xiong S, Xie K, Diao Y, Hong X (2014a) Characterization of the solid electrolyte interphase on lithium anode for preventing the shuttle mechanism in lithium-sulfur batteries. J Power Sources 246:840–845. doi:10.1016/j.jpowsour.2013.08.041
References
187
Xiong S, Xie K, Blomberg E, Jacobsson P, Matic A (2014b) Analysis of the solid electrolyte interphase formed with an ionic liquid electrolyte for lithium-sulfur batteries. J Power Sources 252:150–155. doi:10.1016/j.jpowsour.2013.11.119 Xiong S, Diao Y, Hong X, Chen Y, Xie K (2014c) Characterization of solid electrolyte interphase on lithium electrodes cycled in ether-based electrolytes for lithium batteries. J Electroanal Chem 719:122–126. doi:10.1016/j.jelechem.2014.02.014 Xu W, Xiao J, Zhang J, Wang DY, Zhang JG (2009) Optimization of nonaqueous electrolytes for primary lithium/air batteries operated in ambient environment. J Electrochem Soc 156(10): A773–A779. doi:10.1149/1.3168564 Xu W, Xiao J, Wang DY, Zhang J, Zhang JG (2010) Effects of nonaqueous electrolytes on the performance of lithium/air batteries. J Electrochem Soc 157(2):A219–A224. doi:10.1149/1. 3269928 Xu G, Ding B, Pan J, Nie P, Shen L, Zhang X-W (2014a) High performance lithium-sulfur batteries: advances and challenges. J Mater Chem A 2(32):12662–12676. doi:10.1039/ c4ta02097a Xu W, Wang J, Ding F, Chen X, Nasybulin E, Zhang Y, Zhang J-G (2014b) Lithium metal anodes for rechargeable batteries. Energy Environ Sci 7(2):513–537. doi:10.1039/c3ee40795k Yamin H, Peled E (1983) Electrochemistry of a nonaqueous lithium/sulfur cell. J Power Sources 9:281–287 Yang XH, Xia YY (2010) The effect of oxygen pressures on the electrochemical profile of lithium/oxygen battery. J Solid State Electrochem 14(1):109–114. doi:10.1007/s10008-0090791-8 Yang Y, McDowell MT, Jackson A, Cha JJ, Hong SS, Cui Y (2010) New nanostructured Li2S/silicon rechargeable battery with high specific energy. Nano Lett 10(4):1486–1491. doi:10.1021/nl100504q Yang Y, Zheng G, Cui Y (2013a) Nanostructured sulfur cathodes. Chem Soc Rev 42(7):3018– 3032. doi:10.1039/C2CS35256G Yang Y, Zheng G, Cui Y (2013b) A membrane-free lithium/polysulfide semi-liquid battery for large-scale energy storage. Energy Environ Sci 6(5):1552–1558. doi:10.1039/C3EE00072A Yao H, Yan K, Li W, Zheng G, Kong D, Seh ZW, Narasimhan VK, Liang Z, Cui Y (2014) Improved lithium-sulfur batteries with a conductive coating on the separator to prevent the accumulation of inactive S-related species at the cathode-separator interface. Energy Environ Sci. doi:10.1039/C4EE01377H Ye H, Xu JJ (2008) Polymer electrolytes based on ionic liquids and their application to solid-state thin-film Li-oxygen batteries. ECS Trans 3(42):73–81. doi:10.1149/1.2838194 Yim T, Park M-S, Yu J-S, Kim KJ, Im KY, Kim J-H, Jeong G, Jo YN, Woo S-G, Kang KS, Lee I, Kim Y-J (2013) Effect of chemical reactivity of polysulfide toward carbonate-based electrolyte on the electrochemical performance of Li–S batteries. Electrochim Acta 107:454–460. doi:10. 1016/j.electacta.2013.06.039 Yin Y-X, Xin S, Guo Y-G, Wan L-J (2013) Lithium-sulfur batteries: electrochemistry, materials, and prospects. Angew Chem Int Ed 52(50):13186–13200. doi:10.1002/anie.201304762 Younesi R, Hahlin M, Roberts M, Edström K (2013) The SEI layer formed on lithium metal in the presence of oxygen: a seldom considered component in the development of the Li–O2 battery. J Power Sources 225:40–45. doi:10.1016/j.jpowsour.2012.10.011 Yu X, Xie J, Yang J, Wang K (2004) All solid-state rechargeable lithium cells based on nano-sulfur composite cathodes. J Power Sources 132(1–2):181–186. doi:10.1016/j.jpowsour. 2004.01.034 Yu M, Yuan W, Li C, Hong J-D, Shi G (2014) Performance enhancement of a graphene-sulfur composite as a lithium-sulfur battery electrode by coating with an ultrathin Al2O3 film via atomic layer deposition. J Mater Chem A 2(20):7360–7366. doi:10.1039/C4TA00234B Zaghib K (2012) Lithium metal for rechargeable polymer and metal-air batteries: challenges and opportunities
188
4 Application of Lithium Metal Anodes
Zhang XG (2009) Zinc electrodes: overview. In: Garche J, Dyer C, Moseley P, Ogumi Z, Rand D, Scrosati B (eds) Encyclopeida of electrochemical power sources, vol 5. Elsevier, Amsterdam, pp 454–468 Zhang SS (2012a) Role of LiNO3 in rechargeable lithium/sulfur battery. Electrochim Acta 70:344– 348. doi:10.1016/j.electacta.2012.03.081 Zhang SS (2012b) Effect of discharge cutoff voltage on reversibility of lithium/sulfur batteries with LiNO3-contained electrolyte. J Electrochem Soc 159(7):A920–A923. doi:10.1149/2.002207jes Zhang SS (2013a) Liquid electrolyte lithium/sulfur battery: fundamental chemistry, problems, and solutions. J Power Sources 231:153–162. doi:10.1016/j.jpowsour.2012.12.102 Zhang SS (2013b) A concept for making poly(ethylene oxide) based composite gel polymer electrolyte lithium/sulfur battery. J Electrochem Soc 160(9):A1421–A1424. doi:10.1149/2. 058309jes Zhang SS, Read JA (2012) A new direction for the performance improvement of rechargeable lithium/sulfur batteries. J Power Sources 200:77–82. doi:10.1016/j.jpowsour.2011.10.076 Zhang SS, Tran DT (2013) How a gel polymer electrolyte affects performance of lithium/sulfur batteries. Electrochim Acta 114:296–302. doi:10.1016/j.electacta.2013.10.069 Zhang JG, Wang DY, Xu W, Xiao J, Williford RE (2010) Ambient operation of Li/Air batteries. J Power Sources 195(13):4332–4337. doi:10.1016/j.jpowsour.2010.01.022 Zhang Y, Zhao Y, Bakenov Z, Gosselink D, Chen P (2014) Poly(vinylidene fluoride-co-hexafluoropropylene)/poly(methylmethacrylate)/nanoclay composite gel polymer electrolyte for lithium/sulfur batteries. J Solid State Electrochem 18(4):1111–1116. doi:10. 1007/s10008-013-2366-y Zheng JP, Liang RY, Hendrickson M, Plichta EJ (2008) Theoretical energy density of Li-air batteries. J Electrochem Soc 155(6):A432–A437. doi:10.1149/1.2901961 Zheng J, Lv D, Gu M, Wang C, Zhang J-G, Liu J, Xiao J (2013a) How to obtain reproducible results for lithium sulfur batteries? J Electrochem Soc 160(11):A2288–A2292. doi:10.1149/2. 106311jes Zheng J, Gu M, Chen H, Meduri P, Engelhard MH, Zhang J-G, Liu J, Xiao J (2013b) Ionic liquid-enhanced solid state electrolyte interface (SEI) for lithium-sulfur batteries. J Mater Chem A 1(29):8464–8470. doi:10.1039/C3TA11553D Zheng G, Lee SW, Liang Z, Lee HW, Yan K, Yao H, Wang H, Li W, Chu S, Cui Y (2014) Interconnected hollow carbon nanospheres for stable lithium metal anodes. Nat Nanotechnol 9 (8):618–623. doi:10.1038/nnano.2014.152 Zhou G, Pei S, Li L, Wang D-W, Wang S, Huang K, Yin L-C, Li F, Cheng H-M (2014) A graphene–pure-sulfur sandwich structure for ultrafast, long-life lithium-sulfur batteries. Adv Mater 26(4):625–631. doi:10.1002/adma.201302877 Zu C, Manthiram A (2014) Stabilized lithium-metal surface in a polysulfide-rich environment of lithium-sulfur batteries. J Phys Chem Lett 5(15):2522–2527. doi:10.1021/jz501352e Zu C, Su Y-S, Fu Y, Manthiram A (2013) Improved lithium-sulfur cells with a treated carbon paper interlayer. Phys Chem Chem Phys 15(7):2291–2297. doi:10.1039/C2CP43394J
Chapter 5
Perspectives
Li metal is an ideal anode material for rechargeable batteries, including Li-air, Li-S, and Li metal batteries using intercalation compounds or conversion compounds as cathode materials. Unfortunately, Li dendrite growth and low CE during the charge/discharge processes have largely prevented the use of Li metal as an anode for rechargeable batteries. To date, partial solutions for the prevention of Li dendrite growth have been identified, although they are only effective under certain conditions. For example, the PEO-based block copolymer electrolytes developed by Balsara et al. (2008, 2009) can prevent dendrite growth by using the strong shear strength of block copolymers, but these copolymers must work at elevated temperatures (*80 °C) when Li salt-PEO has acceptable conductivities. Selective electrolyte additives can prevent dendrite growth by a physical blocking mechanism or self-healing mechanism (Matsuda 1993; Aurbach et al. 2002b; Yoon et al. 2008; Aurbach and Chusid (Youngman) 1993), but most of these additives can prevent dendrite growth only at limited current densities. Too large a current density will induce a large voltage drop, which may force the additives to be deposited with Li. Most of these additives will be consumed during the cycling, so their effect will decrease with cycling iterations. On the other hand, improving the CE of Li cycling is a more challenging task. In this regard, the use of selected high concentration electrolytes seems to be a very promising approach that may lead to a CE of more than 99 % without dendrite formation (Qian et al. 2015b, 2016; Yuki Yamada and Yoshitaka Tateyama 2014). Although we are still a long way from realizing a superconcentrated electrolyte in practical application, we believe that this particular electrochemical feature and its detailed mechanisms will be of great value to the development of rechargeable Li metal batteries. Generally speaking, high CE is a highly fundamental criterion for stable cycling of the Li metal anode. To have a high CE, side reactions between freshly/native deposited Li and electrolyte must be minimized. These reactions are proportional to the chemical and electrochemical activity of native Li when they are in direct
© Springer International Publishing Switzerland 2017 J.-G. Zhang et al., Lithium Metal Anodes and Rechargeable Lithium Metal Batteries, Springer Series in Materials Science 249, DOI 10.1007/978-3-319-44054-5_5
189
190
5 Perspectives
contact with surrounding electrolyte. They are also proportional to the surface area of deposited Li. Therefore, a high CE of Li cycling is usually a direct result of low reactivity between freshly deposited Li and the electrolyte, as well as of a low surface area of deposited Li. By contrast, a dendritic Li deposition always has a high surface area. This means that high CE Li deposition/stripping always correlates with low surface area Li deposition and suppressed Li dendrite growth. A stable CE value during long-term cycling also means that the SEI layer formed during Li deposition is relatively stable and that formation of new SEI layers is very minimal during each cycle. On the other hand, some electrolytes can foster dendrite-free Li deposition, but exhibit a CE of less than 80 % (Qian et al. 2015a; Ding et al. 2013). Therefore, the enhancement of CE is a fundamental factor in determining long-term, stable cycling of Li metal anodes. To enable broad application of the Li anode, further fundamental studies need to be conducted to simultaneously address the two barriers discussed above. Future work also needs greater focus on addressing the cause of the problems instead of their consequences. The causes of Li dendrite formation and growth are the Li+ concentration gradient during the charge/discharge process and nonuniform surface deposition. The origin of low CE is thermodynamic incompatibility between fresh Li metal and the electrolyte. The results of modeling and simulations suggest that an electrolyte with high ionic conductivity and a Li+ transference number close to unity can dramatically flatten the concentration gradient, extending the “Sand’s time,” i.e., delaying or even stopping, Li dendrite formation. High shear modulus is also an important property for a functional electrolyte to suppress the Li dendrite growth. In an effort to clearly understand the mechanism of Li dendrite formation and growth, various techniques have been adopted to characterize the morphology, components and structures of Li surfaces. Although some ex situ techniques provide much useful information, Li deposition/stripping is an interfacial reaction, so ex situ characterizations might destroy its original fine morphology. To understand the real dendrite formation and SEI layer formation, as well as their interactions, in situ or operando techniques (including operando TEM, SEM, AFM, or in situ optical techniques with high magnifications) can provide more information about local concentration gradient, Li dynamic nucleation and dendritic growth. They can also be used to observe the formation of dendrites and SEI layers at the same time. This is critical to finding an ultimate solution to overcoming the two barriers that have prevented the application of the Li anode to date. Many modeling studies have been performed and have led to several critical conclusions that guided development of Li metal batteries. However, most of this modeling work has focused on the physical properties of dendrite growth. Most of the significant progress on increasing the CE of Li deposition/stripping [especially Aurbach’s work (Aurbach et al. 1987, 1997, 2002a)] has been achieved by experimental efforts. To enable the application of Li metal anodes, the chemical reactions between Li metal and electrolytes (including solvents, salts and additives)
5 Perspectives
191
need to be modeled and/or simulated to get a better understanding of the formation of the SEI layer and its chemical stability, ionic conductivity, and mechanical flexibility. Furthermore, the interactions between Li dendrites and the SEI layer need to be better understood and incorporated into the simulation process to get a realistic picture of the Li deposition/stripping process. For example, an SEI layer with limited flexibility will be able to tolerate small variations on the Li surface. This will prevent repeated breakdown and reformation of the SEI layer, which is one of the main sources of low CE during a Li anode’s operation. Li is thermodynamically unstable with any organic solvent; they react instantaneously to form an SEI layer. Therefore, the stability and flexibility of the SEI layer is the most critical factor determining the performance of a Li metal anode, especially its CE. The components of the electrolytes, including solvents, Li salts, and additives determine the chemical composition, ionic conductivity, and mechanical properties of the SEI layer. Compared to liquid electrolytes, polymer materials (especially PEO) are much more compatible with fresh Li. Therefore, PEO and its derivatives have been investigated extensively for their applications to rechargeable Li metal batteries over the last three decades. However, the PEO electrolyte still cannot prevent Li dendrite growth because it has a relatively low ionic conductivity and a small Li+ transference number, causing a steep Li+ concentration gradient during polarization, as anticipated by modeling. An effective strategy is to develop single-ion conductors with a unity Li+ transference number, plus a strong shear modulus with block copolymers, which greatly reduce the concentration gradient and effectively retard Li dendrite formation and growth. Another possible solution is to combine the self-healing mechanism with an electrolyte that can form a stable SEI layer with Li. Further improvement in the conductivity of the electrolyte can reduce the internal voltage drop under high current density conditions. Development of inorganic, solid-state Li+ ion conductors with good mechanical strength and stability and high Li ionic conductivity as well as excellent compatibility with Li metal could present an ideal solution for rechargeable Li metal batteries. Creating practical applications of Li metal anodes for rechargeable batteries may require a combination of different approaches. A well balanced Li deposition/stripping rate is also critical for long term stability of Li cycling. It is likely that no single ideal solution will make a Li anode work well under all conditions, so different solutions may have to be designed for different applications. Although there are still many obstacles to overcome, we are optimistic that Li metal can be used as an anode in rechargeable batteries in the near future. This will enable wide applications of rechargeable Li-S batteries, Li-air batteries, and Li metal batteries using intercalation compounds as cathode.
192
5 Perspectives
References Aurbach D, Chusid (Youngman) O (1993) In situ FTIR spectroelectrochemical studies of surface films formed on Li and nonactive electrodes at low potentials in Li salt solutions containing CO2. J Electrochem Soc 140(11):L155–L157 Aurbach D, Daroux ML, Faguy PW, Yeager E (1987) Identification of surface films formed on lithium in propylene carbonate solutions. J Electrochem Soc 134(7):1611–1620 Aurbach D, Zaban A, Ein-Eli Y, Weissman I, Chusid O, Markovsky B, Levi M, Levi E, Schechter A, Granot E (1997) Recent studies on the correlation between surface chemistry, morphology, three-dimensional structures and performance of Li and Li-C intercalation anodes in several important electrolyte systems. J Power Sources 68:91–98 Aurbach D, Zinigrad E, Cohen Y, Teller H (2002a) A short review of failure mechanisms of lithium metal and lithiated graphite anodes in liquid electrolyte solutions. Solid State Ionics 148:405–416 Aurbach D, Zinigrad E, Teller H, Cohen Y, Salitra G, Yamin H, Dan P, Elster E (2002b) Attempts to improve the behavior of Li electrodes in rechargeable lithium batteries. J Electrochem Soc 149(10):A1267–A1277. doi:10.1149/1.1502684 Balsara N (2008). http://www.1.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2008/ exploratory_battery/merit08_balsarapdf Balsara NP, Singh M, Eitouni HB, Gomez ED (2009). US Patent Appl No 0263725 A1 Ding F, Xu W, Graff GL, Zhang J, Sushko ML, Chen X, Shao Y, Engelhard MH, Nie Z, Xiao J, Liu X, Sushko PV, Liu J, Zhang J-G (2013) Dendrite-free lithium deposition via self-healing electrostatic shield mechanism. J Am Chem Soc 135(11):4450–4456. doi:10.1021/ja312241y Matsuda Y (1993) Behavior of lithium/electrolyte interface in organic solutions. J Power Sources 43–44:1–7 Qian J, Xu W, Bhattacharya P, Engelhard M, Henderson WA, Zhang Y, Zhang J-G (2015a) Dendrite-free Li deposition using trace-amounts of water as an electrolyte additive. Nano Energy 15:135–144. doi:10.1016/j.nanoen.2015.04.009 Qian J, Henderson WA, Xu W, Bhattacharya P, Engelhard M, Borodin O, Zhang JG (2015b) High rate and stable cycling of lithium metal anode. Nat Commun 6:6362. doi:10.1038/ ncomms7362 Qian JF, Adams BD, Zheng JM, Xu W, Henderson WA, Wang J, Bowden ME, Xu SC, Hu JZ, Zhang JZ (2016) Anode-free rechargeable lithium metal batteries. Adv Funct Mater 2016. doi:10.1002/adfm.201602353 Yoon S, Lee J, Kim S-O, Sohn H-J (2008) Enhanced cyclability and surface characteristics of lithium batteries by Li–Mg co-deposition and addition of HF acid in electrolyte. Electrochim Acta 53(5):2501–2506. doi:10.1016/j.electacta.2007.10.019 Yamada Y, Furukawa K, Sodeyama K, Kikuchi K, Yaegashi M, Tateyama Y, Yamada A (2014) J Am Chem Soc 136:5039–5046
Index
A Aligned, 3, 24, 45, 68 Alloys, 97, 98 Amorphous, 83–85 Anode, 3, 6, 35, 75, 79, 86, 93, 95, 127, 153, 154, 156, 158, 160, 164, 167, 172, 174, 176–179, 189–191 Anode-free batteries, 178, 179 Aqueous electrolyte, 92 B Blocking, 2, 161, 162 C Carbonate, 13, 18, 20, 51, 52, 68, 164 Charge, 6, 79, 82, 89, 106, 109, 111, 125, 154, 158, 166, 172, 176 Concentrated electrolyte, 50, 73, 75, 78, 126, 165, 179 Contaminants, 10, 48 Corrosion, 100, 156, 160, 163, 165, 173, 175 Coulombic efficiency, 1 Cross section, 6, 8 Crystalline, 20, 58, 75 Current density, 8, 21, 28, 29, 33, 34, 36, 105, 109, 111, 113, 189 Cycle life, 1, 92, 111, 115, 157, 158, 178 Cyclic voltammetry, 72 D Dendrite, 2, 5, 11, 21, 26, 29, 31, 33–35, 86, 91, 113, 123, 126, 163, 170, 190 Deposition, 2, 5, 10, 23, 24, 30, 48, 70, 96, 103, 113, 127, 190 Dilute, 58, 76, 77 Discharge, 86, 89, 91, 111, 115, 116, 158, 167, 174, 176 Dissolution, 13, 165
E Electrochemical, 1, 2, 8, 10, 13, 29, 34, 48, 72, 92, 93, 95, 96, 111, 127 Electrolyte additives, 167, 189 Ether, 179 F Figure-of-merit, 47 I Ionic liquid, 72 L LiFSI, 73, 79, 179 Li metal batteries, 1, 3, 75, 153, 160, 189–191 LiPON, 91, 179 LISICO glass, 92 Lithium, 1, 18, 76, 154, 163 Lithium salts, 17, 76 M Modeling, 3, 190 Morphology, 1, 3, 5, 9, 16, 20, 31, 62, 70, 103, 105, 111, 113, 119, 122, 124, 126, 190 Mossy Li, 6, 27 Multi-layer, 48 N Nanorods, 45 Nanosphere, 83, 85 Nernst equation, 80 NMR, 14, 19 Nonaqueous electrolyte, 10, 34, 92, 173, 175 O Optical images, 88 Optical microscopy, 5, 8, 10, 121
© Springer International Publishing Switzerland 2017 J.-G. Zhang et al., Lithium Metal Anodes and Rechargeable Lithium Metal Batteries, Springer Series in Materials Science 249, DOI 10.1007/978-3-319-44054-5
193
194 P Polysulfides, 158–160, 162, 164, 167 Pulsed plating, 105, 127 R Rechargeable batteries, 1, 189, 191 Roughness, 36, 103, 105, 127 S Safety, 5, 45, 94, 155, 156, 167 Sand’s time, 34, 35, 190 Self-healing electrostatic shield, 79 SEM, 6–8, 10, 21, 22, 68, 85, 124, 168, 190 Separator, 31, 86, 124–126, 160, 161, 178 Short circuit, 20, 33, 74, 91, 96, 114 Side reaction, 2, 45, 96, 127, 153, 160, 165
Index Solid electrolyte interphase, 2, 85 Solid polymer electrolyte, 86, 171 Stack pressure, 123–126 Storage time, 118 Stripping, 2, 3, 8, 10, 13, 45, 63, 89, 105, 114, 127 Substrate, 24, 31, 32, 81, 85, 92, 96, 127 Super-concentrated, 189 Superconcentrated, 76–78 Surface coating, 48, 68, 69 Surface morphology, 20, 24, 167 T TEM, 11, 13 Temperature, 21, 24, 51, 59, 72, 85, 88, 94, 95, 97, 98, 118, 119, 159, 160, 176